{smcl}
{com}{sf}{ul off}{txt}{.-}
      name:  {res}<unnamed>
       {txt}log:  {res}C:\Users\MB\Desktop\JOP_Replicate\ReplicationUpload\LogforRegressionModels.smcl
  {txt}log type:  {res}smcl
 {txt}opened on:  {res} 9 Feb 2023, 18:56:53

{com}. do "C:\Users\MB\Desktop\JOP_Replicate\ReplicationUpload\ReplicationFile.do"
{txt}
{com}. *use "BCG_JOP_ReplicationData.dta" 
. 
. xtset audience year, yearly
{res}
{col 1}{txt:Panel variable: }{res:audience}{txt: (unbalanced)}
{p 1 16 2}{txt:Time variable: }{res:year_a}{txt:, }{res:{bind:1939}}{txt: to }{res:{bind:2018}}{txt:, but with gaps}{p_end}
{txt}{col 10}Delta: {res}1 year
{txt}
{com}. 
. 
. 
. *** MAIN PAPER TABLE 2***
.                 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-185.77755}  
Iteration 1:{space 3}log pseudolikelihood = {res:-177.95976}  
Iteration 2:{space 3}log pseudolikelihood = {res: -175.0278}  
Iteration 3:{space 3}log pseudolikelihood = {res:-174.99552}  
Iteration 4:{space 3}log pseudolikelihood = {res:-174.99548}  
Iteration 5:{space 3}log pseudolikelihood = {res:-174.99548}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,413}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:23.75}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-174.99548}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0580}

{txt}{ralign 86:(Std. err. adjusted for {res:140} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .7032983{col 34}{space 2} .2362836{col 45}{space 1}    2.98{col 54}{space 3}0.003{col 62}{space 4} .2401908{col 75}{space 3} 1.166406
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5120046{col 34}{space 2} .2168771{col 45}{space 1}   -2.36{col 54}{space 3}0.018{col 62}{space 4}-.9370758{col 75}{space 3}-.0869333
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.2578904{col 34}{space 2} .2573307{col 45}{space 1}   -1.00{col 54}{space 3}0.316{col 62}{space 4}-.7622492{col 75}{space 3} .2464685
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.218854{col 34}{space 2} .3935801{col 45}{space 1}  -13.26{col 54}{space 3}0.000{col 62}{space 4}-5.990257{col 75}{space 3}-4.447452
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-121.55265}  
Iteration 1:{space 3}log pseudolikelihood = {res:-115.75851}  
Iteration 2:{space 3}log pseudolikelihood = {res:-113.26616}  
Iteration 3:{space 3}log pseudolikelihood = {res:-113.09187}  
Iteration 4:{space 3}log pseudolikelihood = {res:-113.09043}  
Iteration 5:{space 3}log pseudolikelihood = {res:-113.09043}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,681}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:8.88}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0309}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-113.09043}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0696}

{txt}{ralign 86:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .6062375{col 34}{space 2} .3571621{col 45}{space 1}    1.70{col 54}{space 3}0.090{col 62}{space 4}-.0937874{col 75}{space 3} 1.306262
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.262755{col 34}{space 2} .4930823{col 45}{space 1}   -2.56{col 54}{space 3}0.010{col 62}{space 4}-2.229178{col 75}{space 3}-.2963309
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .2924625{col 34}{space 2} .3421655{col 45}{space 1}    0.85{col 54}{space 3}0.393{col 62}{space 4}-.3781696{col 75}{space 3} .9630947
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.800323{col 34}{space 2} .5771163{col 45}{space 1}  -10.05{col 54}{space 3}0.000{col 62}{space 4}-6.931451{col 75}{space 3}-4.669196
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-170.03704}  
Iteration 1:{space 3}log pseudolikelihood = {res:-162.32642}  
Iteration 2:{space 3}log pseudolikelihood = {res:-156.64041}  
Iteration 3:{space 3}log pseudolikelihood = {res:-156.62317}  
Iteration 4:{space 3}log pseudolikelihood = {res:-156.62316}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,410}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:35.49}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-156.62316}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0789}

{txt}{ralign 86:(Std. err. adjusted for {res:140} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .9131142{col 34}{space 2} .2522408{col 45}{space 1}    3.62{col 54}{space 3}0.000{col 62}{space 4} .4187314{col 75}{space 3} 1.407497
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}  -.48066{col 34}{space 2} .2339682{col 45}{space 1}   -2.05{col 54}{space 3}0.040{col 62}{space 4}-.9392292{col 75}{space 3}-.0220908
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -.406819{col 34}{space 2} .3265948{col 45}{space 1}   -1.25{col 54}{space 3}0.213{col 62}{space 4}-1.046933{col 75}{space 3}  .233295
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.512878{col 34}{space 2}  .434922{col 45}{space 1}  -12.68{col 54}{space 3}0.000{col 62}{space 4}-6.365309{col 75}{space 3}-4.660446
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>1945 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-257.69965}  
Iteration 1:{space 3}log pseudolikelihood = {res: -245.2775}  
Iteration 2:{space 3}log pseudolikelihood = {res:-240.80828}  
Iteration 3:{space 3}log pseudolikelihood = {res:-240.74026}  
Iteration 4:{space 3}log pseudolikelihood = {res: -240.7402}  
Iteration 5:{space 3}log pseudolikelihood = {res: -240.7402}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,681}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:35.21}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-240.7402}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0658}

{txt}{ralign 86:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .6987064{col 34}{space 2} .2220066{col 45}{space 1}    3.15{col 54}{space 3}0.002{col 62}{space 4} .2635815{col 75}{space 3} 1.133831
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.6727956{col 34}{space 2} .2299817{col 45}{space 1}   -2.93{col 54}{space 3}0.003{col 62}{space 4}-1.123551{col 75}{space 3}-.2220398
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -.113742{col 34}{space 2} .2092963{col 45}{space 1}   -0.54{col 54}{space 3}0.587{col 62}{space 4}-.5239552{col 75}{space 3} .2964713
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-4.924702{col 34}{space 2} .3589409{col 45}{space 1}  -13.72{col 54}{space 3}0.000{col 62}{space 4}-5.628213{col 75}{space 3} -4.22119
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-97.941224}  
Iteration 1:{space 3}log pseudolikelihood = {res:-91.890155}  
Iteration 2:{space 3}log pseudolikelihood = {res:-89.851029}  
Iteration 3:{space 3}log pseudolikelihood = {res:-89.844179}  
Iteration 4:{space 3}log pseudolikelihood = {res:-89.844178}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:26.69}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-89.844178}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0827}

{txt}{ralign 86:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-1.028042{col 34}{space 2} .2722958{col 45}{space 1}   -3.78{col 54}{space 3}0.000{col 62}{space 4}-1.561732{col 75}{space 3} -.494352
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3181166{col 34}{space 2} .1188786{col 45}{space 1}    2.68{col 54}{space 3}0.007{col 62}{space 4} .0851189{col 75}{space 3} .5511143
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.9699223{col 34}{space 2} .4257929{col 45}{space 1}   -2.28{col 54}{space 3}0.023{col 62}{space 4}-1.804461{col 75}{space 3}-.1353836
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-1.930569{col 34}{space 2} .3243576{col 45}{space 1}   -5.95{col 54}{space 3}0.000{col 62}{space 4}-2.566299{col 75}{space 3} -1.29484
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res: -95.02947}  
Iteration 1:{space 3}log pseudolikelihood = {res:-89.268655}  
Iteration 2:{space 3}log pseudolikelihood = {res:-86.851685}  
Iteration 3:{space 3}log pseudolikelihood = {res:-86.842892}  
Iteration 4:{space 3}log pseudolikelihood = {res: -86.84289}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:25.28}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-86.84289}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0861}

{txt}{ralign 86:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-1.065486{col 34}{space 2} .2921816{col 45}{space 1}   -3.65{col 54}{space 3}0.000{col 62}{space 4}-1.638151{col 75}{space 3}-.4928206
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3400882{col 34}{space 2} .1214351{col 45}{space 1}    2.80{col 54}{space 3}0.005{col 62}{space 4} .1020797{col 75}{space 3} .5780967
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -.935321{col 34}{space 2} .4292657{col 45}{space 1}   -2.18{col 54}{space 3}0.029{col 62}{space 4}-1.776666{col 75}{space 3}-.0939756
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -1.97955{col 34}{space 2} .3414956{col 45}{space 1}   -5.80{col 54}{space 3}0.000{col 62}{space 4}-2.648869{col 75}{space 3} -1.31023
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>1945, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     5,880
                                                {txt}F(3, 142)         =  {res}     4.72
                                                {txt}Prob > F          = {res}    0.0036
                                                {txt}R-squared         = {res}    0.0030
                                                {txt}Root MSE          =    {res} .10972

{txt}{ralign 86:(Std. err. adjusted for {res:143} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0071277{col 34}{space 2} .0029411{col 45}{space 1}    2.42{col 54}{space 3}0.017{col 62}{space 4} .0013138{col 75}{space 3} .0129416
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0027069{col 34}{space 2} .0012072{col 45}{space 1}   -2.24{col 54}{space 3}0.026{col 62}{space 4}-.0050934{col 75}{space 3}-.0003204
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0014958{col 34}{space 2} .0014893{col 45}{space 1}    1.00{col 54}{space 3}0.317{col 62}{space 4}-.0014483{col 75}{space 3} .0044399
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}  .000155{col 34}{space 2} .0032778{col 45}{space 1}    0.05{col 54}{space 3}0.962{col 62}{space 4}-.0063246{col 75}{space 3} .0066345
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. *** MAIN PAPER TABLE 3 ***
. 
. *Program (Column 1)
. logit audience_start tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-134.04454}  
Iteration 1:{space 3}log pseudolikelihood = {res:-124.34538}  
Iteration 2:{space 3}log pseudolikelihood = {res:-109.60556}  
Iteration 3:{space 3}log pseudolikelihood = {res:-107.88078}  
Iteration 4:{space 3}log pseudolikelihood = {res:-107.81225}  
Iteration 5:{space 3}log pseudolikelihood = {res:-107.81203}  
Iteration 6:{space 3}log pseudolikelihood = {res:-107.81203}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,582}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:116.83}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-107.81203}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1957}

{txt}{ralign 86:(Std. err. adjusted for {res:122} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .6748358{col 34}{space 2} .3511436{col 45}{space 1}    1.92{col 54}{space 3}0.055{col 62}{space 4} -.013393{col 75}{space 3} 1.363065
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5146111{col 34}{space 2} .3303833{col 45}{space 1}   -1.56{col 54}{space 3}0.119{col 62}{space 4} -1.16215{col 75}{space 3} .1329282
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.3490265{col 34}{space 2} .3367678{col 45}{space 1}   -1.04{col 54}{space 3}0.300{col 62}{space 4}-1.009079{col 75}{space 3} .3110261
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .3913616{col 34}{space 2} .1233764{col 45}{space 1}    3.17{col 54}{space 3}0.002{col 62}{space 4} .1495484{col 75}{space 3} .6331748
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2}-.0523419{col 34}{space 2} .0991869{col 45}{space 1}   -0.53{col 54}{space 3}0.598{col 62}{space 4}-.2467446{col 75}{space 3} .1420607
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} 2.390978{col 34}{space 2} .4419407{col 45}{space 1}    5.41{col 54}{space 3}0.000{col 62}{space 4}  1.52479{col 75}{space 3} 3.257165
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} -.019304{col 34}{space 2} .0266377{col 45}{space 1}   -0.72{col 54}{space 3}0.469{col 62}{space 4}-.0715129{col 75}{space 3} .0329049
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} -.040671{col 34}{space 2} .0432384{col 45}{space 1}   -0.94{col 54}{space 3}0.347{col 62}{space 4}-.1254168{col 75}{space 3} .0440747
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.532657{col 34}{space 2} .4881487{col 45}{space 1}  -11.33{col 54}{space 3}0.000{col 62}{space 4}-6.489411{col 75}{space 3}-4.575903
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-89.841707}  
Iteration 1:{space 3}log pseudolikelihood = {res:-77.359062}  
Iteration 2:{space 3}log pseudolikelihood = {res:-70.710561}  
Iteration 3:{space 3}log pseudolikelihood = {res:-65.710847}  
Iteration 4:{space 3}log pseudolikelihood = {res:-64.975469}  
Iteration 5:{space 3}log pseudolikelihood = {res: -64.90701}  
Iteration 6:{space 3}log pseudolikelihood = {res:-64.906656}  
Iteration 7:{space 3}log pseudolikelihood = {res:-64.906656}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,804}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:102.80}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-64.906656}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.2775}

{txt}{ralign 86:(Std. err. adjusted for {res:125} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .1041714{col 34}{space 2} .3211129{col 45}{space 1}    0.32{col 54}{space 3}0.746{col 62}{space 4}-.5251983{col 75}{space 3} .7335412
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-2.024996{col 34}{space 2} .9120093{col 45}{space 1}   -2.22{col 54}{space 3}0.026{col 62}{space 4}-3.812502{col 75}{space 3}-.2374909
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.1229451{col 34}{space 2} .4453547{col 45}{space 1}   -0.28{col 54}{space 3}0.783{col 62}{space 4}-.9958243{col 75}{space 3} .7499342
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .3423966{col 34}{space 2} .3246744{col 45}{space 1}    1.05{col 54}{space 3}0.292{col 62}{space 4}-.2939536{col 75}{space 3} .9787467
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2} .0448463{col 34}{space 2} .0972813{col 45}{space 1}    0.46{col 54}{space 3}0.645{col 62}{space 4}-.1458215{col 75}{space 3} .2355141
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} 2.917138{col 34}{space 2} .8712816{col 45}{space 1}    3.35{col 54}{space 3}0.001{col 62}{space 4} 1.209458{col 75}{space 3} 4.624818
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2}-.1072276{col 34}{space 2} .0579825{col 45}{space 1}   -1.85{col 54}{space 3}0.064{col 62}{space 4}-.2208713{col 75}{space 3} .0064161
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} .0927484{col 34}{space 2} .0455736{col 45}{space 1}    2.04{col 54}{space 3}0.042{col 62}{space 4} .0034257{col 75}{space 3} .1820711
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.405631{col 34}{space 2} .6230536{col 45}{space 1}   -8.68{col 54}{space 3}0.000{col 62}{space 4}-6.626794{col 75}{space 3}-4.184469
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{p 0 6 2}Note: 23 failures and 0 successes completely determined.{p_end}

{com}. 
. *Explore (Column 3)
. logit explore_start tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-128.63508}  
Iteration 1:{space 3}log pseudolikelihood = {res:-120.06619}  
Iteration 2:{space 3}log pseudolikelihood = {res:-102.80086}  
Iteration 3:{space 3}log pseudolikelihood = {res:-100.32866}  
Iteration 4:{space 3}log pseudolikelihood = {res:-100.16424}  
Iteration 5:{space 3}log pseudolikelihood = {res:-100.16381}  
Iteration 6:{space 3}log pseudolikelihood = {res:-100.16381}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,581}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:100.36}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-100.16381}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.2213}

{txt}{ralign 86:(Std. err. adjusted for {res:122} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .9048131{col 34}{space 2} .3685129{col 45}{space 1}    2.46{col 54}{space 3}0.014{col 62}{space 4}  .182541{col 75}{space 3} 1.627085
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5257582{col 34}{space 2} .3604532{col 45}{space 1}   -1.46{col 54}{space 3}0.145{col 62}{space 4}-1.232234{col 75}{space 3} .1807172
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.4864544{col 34}{space 2} .3982621{col 45}{space 1}   -1.22{col 54}{space 3}0.222{col 62}{space 4}-1.267034{col 75}{space 3} .2941249
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .3629091{col 34}{space 2} .1335034{col 45}{space 1}    2.72{col 54}{space 3}0.007{col 62}{space 4} .1012473{col 75}{space 3}  .624571
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2} .0121005{col 34}{space 2} .0897006{col 45}{space 1}    0.13{col 54}{space 3}0.893{col 62}{space 4}-.1637095{col 75}{space 3} .1879104
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} 2.603901{col 34}{space 2} .5808704{col 45}{space 1}    4.48{col 54}{space 3}0.000{col 62}{space 4} 1.465416{col 75}{space 3} 3.742386
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} -.028909{col 34}{space 2} .0291138{col 45}{space 1}   -0.99{col 54}{space 3}0.321{col 62}{space 4} -.085971{col 75}{space 3}  .028153
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2}-.0197238{col 34}{space 2} .0460425{col 45}{space 1}   -0.43{col 54}{space 3}0.668{col 62}{space 4}-.1099654{col 75}{space 3} .0705177
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.790674{col 34}{space 2} .5455201{col 45}{space 1}  -10.61{col 54}{space 3}0.000{col 62}{space 4}-6.859873{col 75}{space 3}-4.721474
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & year_a>1945 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-192.26017}  
Iteration 1:{space 3}log pseudolikelihood = {res:-173.78266}  
Iteration 2:{space 3}log pseudolikelihood = {res:-154.60546}  
Iteration 3:{space 3}log pseudolikelihood = {res:-152.52894}  
Iteration 4:{space 3}log pseudolikelihood = {res:-152.46768}  
Iteration 5:{space 3}log pseudolikelihood = {res:-152.46743}  
Iteration 6:{space 3}log pseudolikelihood = {res:-152.46743}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,804}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:127.38}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-152.46743}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.2070}

{txt}{ralign 86:(Std. err. adjusted for {res:125} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .4978753{col 34}{space 2}  .265406{col 45}{space 1}    1.88{col 54}{space 3}0.061{col 62}{space 4}-.0223109{col 75}{space 3} 1.018062
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.7293115{col 34}{space 2} .3752579{col 45}{space 1}   -1.94{col 54}{space 3}0.052{col 62}{space 4}-1.464803{col 75}{space 3} .0061804
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.3082736{col 34}{space 2}  .267053{col 45}{space 1}   -1.15{col 54}{space 3}0.248{col 62}{space 4}-.8316878{col 75}{space 3} .2151405
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .2751047{col 34}{space 2}  .138968{col 45}{space 1}    1.98{col 54}{space 3}0.048{col 62}{space 4} .0027324{col 75}{space 3}  .547477
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2} .0154643{col 34}{space 2} .0733724{col 45}{space 1}    0.21{col 54}{space 3}0.833{col 62}{space 4} -.128343{col 75}{space 3} .1592715
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} 2.278776{col 34}{space 2} .4942322{col 45}{space 1}    4.61{col 54}{space 3}0.000{col 62}{space 4} 1.310099{col 75}{space 3} 3.247453
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2}-.0442691{col 34}{space 2} .0306455{col 45}{space 1}   -1.44{col 54}{space 3}0.149{col 62}{space 4} -.104333{col 75}{space 3} .0157949
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} .0022134{col 34}{space 2} .0341328{col 45}{space 1}    0.06{col 54}{space 3}0.948{col 62}{space 4}-.0646856{col 75}{space 3} .0691124
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-4.913528{col 34}{space 2} .4333004{col 45}{space 1}  -11.34{col 54}{space 3}0.000{col 62}{space 4}-5.762781{col 75}{space 3}-4.064275
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>1945, vce(cluster audience)

{txt}note: {bf:inttot_a} != 0 predicts failure perfectly;
      {bf:inttot_a} omitted and 41 obs not used.

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-73.547882}  
Iteration 1:{space 3}log pseudolikelihood = {res:-69.735729}  
Iteration 2:{space 3}log pseudolikelihood = {res:-65.750814}  
Iteration 3:{space 3}log pseudolikelihood = {res: -65.68966}  
Iteration 4:{space 3}log pseudolikelihood = {res: -65.68956}  
Iteration 5:{space 3}log pseudolikelihood = {res: -65.68956}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:301}
{txt}{col 57}{lalign 13:Wald chi2({res:7})}{col 70} = {res}{ralign 6:29.64}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0001}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-65.68956}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1068}

{txt}{ralign 86:(Std. err. adjusted for {res:23} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.6927651{col 34}{space 2} .3365439{col 45}{space 1}   -2.06{col 54}{space 3}0.040{col 62}{space 4}-1.352379{col 75}{space 3}-.0331512
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .4749007{col 34}{space 2} .1288298{col 45}{space 1}    3.69{col 54}{space 3}0.000{col 62}{space 4} .2223989{col 75}{space 3} .7274024
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.3986256{col 34}{space 2} .4867528{col 45}{space 1}   -0.82{col 54}{space 3}0.413{col 62}{space 4}-1.352644{col 75}{space 3} .5553924
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2}        0{col 34}{txt}  (omitted)
{space 12}civtot_a {c |}{col 22}{res}{space 2} .1755146{col 34}{space 2} .2387515{col 45}{space 1}    0.74{col 54}{space 3}0.462{col 62}{space 4}-.2924298{col 75}{space 3}  .643459
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} .4750964{col 34}{space 2} .6397857{col 45}{space 1}    0.74{col 54}{space 3}0.458{col 62}{space 4}-.7788606{col 75}{space 3} 1.729053
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} .0927757{col 34}{space 2} .0366631{col 45}{space 1}    2.53{col 54}{space 3}0.011{col 62}{space 4} .0209174{col 75}{space 3}  .164634
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} .0272906{col 34}{space 2} .0334625{col 45}{space 1}    0.82{col 54}{space 3}0.415{col 62}{space 4}-.0382948{col 75}{space 3} .0928759
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-3.319197{col 34}{space 2} .7274564{col 45}{space 1}   -4.56{col 54}{space 3}0.000{col 62}{space 4}-4.744985{col 75}{space 3}-1.893408
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>1945, vce(cluster audience)

{txt}note: {bf:inttot_a} != 0 predicts failure perfectly;
      {bf:inttot_a} omitted and 41 obs not used.

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-70.878055}  
Iteration 1:{space 3}log pseudolikelihood = {res:-67.449222}  
Iteration 2:{space 3}log pseudolikelihood = {res:-62.717215}  
Iteration 3:{space 3}log pseudolikelihood = {res:-62.598834}  
Iteration 4:{space 3}log pseudolikelihood = {res:-62.598627}  
Iteration 5:{space 3}log pseudolikelihood = {res:-62.598627}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:301}
{txt}{col 57}{lalign 13:Wald chi2({res:7})}{col 70} = {res}{ralign 6:27.24}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0003}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-62.598627}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1168}

{txt}{ralign 86:(Std. err. adjusted for {res:23} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.7386823{col 34}{space 2} .3645776{col 45}{space 1}   -2.03{col 54}{space 3}0.043{col 62}{space 4}-1.453241{col 75}{space 3}-.0241234
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .5087829{col 34}{space 2} .1372727{col 45}{space 1}    3.71{col 54}{space 3}0.000{col 62}{space 4} .2397334{col 75}{space 3} .7778324
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.3296881{col 34}{space 2}  .499574{col 45}{space 1}   -0.66{col 54}{space 3}0.509{col 62}{space 4}-1.308835{col 75}{space 3} .6494589
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2}        0{col 34}{txt}  (omitted)
{space 12}civtot_a {c |}{col 22}{res}{space 2} .1681328{col 34}{space 2} .2634259{col 45}{space 1}    0.64{col 54}{space 3}0.523{col 62}{space 4}-.3481724{col 75}{space 3} .6844379
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} .2516474{col 34}{space 2} .6334578{col 45}{space 1}    0.40{col 54}{space 3}0.691{col 62}{space 4} -.989907{col 75}{space 3} 1.493202
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} .1095202{col 34}{space 2} .0397029{col 45}{space 1}    2.76{col 54}{space 3}0.006{col 62}{space 4}  .031704{col 75}{space 3} .1873364
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} .0126297{col 34}{space 2} .0372971{col 45}{space 1}    0.34{col 54}{space 3}0.735{col 62}{space 4}-.0604714{col 75}{space 3} .0857307
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-3.440888{col 34}{space 2} .7723729{col 45}{space 1}   -4.45{col 54}{space 3}0.000{col 62}{space 4}-4.954711{col 75}{space 3}-1.927065
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a if aec2==1 & year_a>1945, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     4,923
                                                {txt}F(8, 125)         =  {res}     3.01
                                                {txt}Prob > F          = {res}    0.0040
                                                {txt}R-squared         = {res}    0.0090
                                                {txt}Root MSE          =    {res} .10434

{txt}{ralign 86:(Std. err. adjusted for {res:126} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0032406{col 34}{space 2} .0026885{col 45}{space 1}    1.21{col 54}{space 3}0.230{col 62}{space 4}-.0020803{col 75}{space 3} .0085616
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0026664{col 34}{space 2} .0014423{col 45}{space 1}   -1.85{col 54}{space 3}0.067{col 62}{space 4}-.0055209{col 75}{space 3}  .000188
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0004424{col 34}{space 2} .0013366{col 45}{space 1}    0.33{col 54}{space 3}0.741{col 62}{space 4}-.0022028{col 75}{space 3} .0030877
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .0151349{col 34}{space 2} .0068894{col 45}{space 1}    2.20{col 54}{space 3}0.030{col 62}{space 4} .0014998{col 75}{space 3} .0287699
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2}-.0011617{col 34}{space 2} .0008928{col 45}{space 1}   -1.30{col 54}{space 3}0.196{col 62}{space 4}-.0029286{col 75}{space 3} .0006053
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} .0051734{col 34}{space 2} .0088335{col 45}{space 1}    0.59{col 54}{space 3}0.559{col 62}{space 4}-.0123092{col 75}{space 3}  .022656
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} -.000038{col 34}{space 2} .0000566{col 45}{space 1}   -0.67{col 54}{space 3}0.503{col 62}{space 4}  -.00015{col 75}{space 3}  .000074
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2}-.0002256{col 34}{space 2} .0001958{col 45}{space 1}   -1.15{col 54}{space 3}0.252{col 62}{space 4}-.0006131{col 75}{space 3}  .000162
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} .0032643{col 34}{space 2} .0036215{col 45}{space 1}    0.90{col 54}{space 3}0.369{col 62}{space 4}-.0039031{col 75}{space 3} .0104316
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. 
. 
. 
. *** APPENDIX TABLE 7 ***
. 
. *Unweighted (Column 1)
. logit audience_start tol_end_total_full_5 Nat_total_full_5 dll_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-134.04454}  
Iteration 1:{space 3}log pseudolikelihood = {res:-124.34538}  
Iteration 2:{space 3}log pseudolikelihood = {res:-109.60556}  
Iteration 3:{space 3}log pseudolikelihood = {res:-107.88078}  
Iteration 4:{space 3}log pseudolikelihood = {res:-107.81225}  
Iteration 5:{space 3}log pseudolikelihood = {res:-107.81203}  
Iteration 6:{space 3}log pseudolikelihood = {res:-107.81203}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,582}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:116.83}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-107.81203}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1957}

{txt}{ralign 86:(Std. err. adjusted for {res:122} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .6748358{col 34}{space 2} .3511436{col 45}{space 1}    1.92{col 54}{space 3}0.055{col 62}{space 4} -.013393{col 75}{space 3} 1.363065
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5146111{col 34}{space 2} .3303833{col 45}{space 1}   -1.56{col 54}{space 3}0.119{col 62}{space 4} -1.16215{col 75}{space 3} .1329282
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.3490265{col 34}{space 2} .3367678{col 45}{space 1}   -1.04{col 54}{space 3}0.300{col 62}{space 4}-1.009079{col 75}{space 3} .3110261
{txt}{space 12}inttot_a {c |}{col 22}{res}{space 2} .3913616{col 34}{space 2} .1233764{col 45}{space 1}    3.17{col 54}{space 3}0.002{col 62}{space 4} .1495484{col 75}{space 3} .6331748
{txt}{space 12}civtot_a {c |}{col 22}{res}{space 2}-.0523419{col 34}{space 2} .0991869{col 45}{space 1}   -0.53{col 54}{space 3}0.598{col 62}{space 4}-.2467446{col 75}{space 3} .1420607
{txt}{space 8}rival_weapon {c |}{col 22}{res}{space 2} 2.390978{col 34}{space 2} .4419407{col 45}{space 1}    5.41{col 54}{space 3}0.000{col 62}{space 4}  1.52479{col 75}{space 3} 3.257165
{txt}{space 14}gdp_pc {c |}{col 22}{res}{space 2} -.019304{col 34}{space 2} .0266377{col 45}{space 1}   -0.72{col 54}{space 3}0.469{col 62}{space 4}-.0715129{col 75}{space 3} .0329049
{txt}{space 11}polity2_a {c |}{col 22}{res}{space 2} -.040671{col 34}{space 2} .0432384{col 45}{space 1}   -0.94{col 54}{space 3}0.347{col 62}{space 4}-.1254168{col 75}{space 3} .0440747
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.532657{col 34}{space 2} .4881487{col 45}{space 1}  -11.33{col 54}{space 3}0.000{col 62}{space 4}-6.489411{col 75}{space 3}-4.575903
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. 
. *Time Weighted (Column 2)
. logit audience_start tol_end_tw_total_full_5 Nat_tw_total_full_5 dll_tw_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-134.04454}  
Iteration 1:{space 3}log pseudolikelihood = {res:-124.15409}  
Iteration 2:{space 3}log pseudolikelihood = {res:-109.63705}  
Iteration 3:{space 3}log pseudolikelihood = {res:-108.05852}  
Iteration 4:{space 3}log pseudolikelihood = {res:-108.00729}  
Iteration 5:{space 3}log pseudolikelihood = {res:-108.00716}  
Iteration 6:{space 3}log pseudolikelihood = {res:-108.00716}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,582}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:125.08}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-108.00716}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1942}

{txt}{ralign 89:(Std. err. adjusted for {res:122} clusters in {res:audience})}
{hline 24}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 25}{c |}{col 37}    Robust
{col 1}         audience_start{col 25}{c |} Coefficient{col 37}  std. err.{col 49}      z{col 57}   P>|z|{col 65}     [95% con{col 78}f. interval]
{hline 24}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_tw_total_full_5 {c |}{col 25}{res}{space 2} 1.237773{col 37}{space 2} .5125352{col 48}{space 1}    2.42{col 57}{space 3}0.016{col 65}{space 4} .2332224{col 78}{space 3} 2.242323
{txt}{space 4}Nat_tw_total_full_5 {c |}{col 25}{res}{space 2}-.6057802{col 37}{space 2} .3755488{col 48}{space 1}   -1.61{col 57}{space 3}0.107{col 65}{space 4}-1.341842{col 78}{space 3}  .130282
{txt}{space 4}dll_tw_total_full_5 {c |}{col 25}{res}{space 2}-.0871463{col 37}{space 2} .4521854{col 48}{space 1}   -0.19{col 57}{space 3}0.847{col 65}{space 4}-.9734133{col 78}{space 3} .7991208
{txt}{space 15}inttot_a {c |}{col 25}{res}{space 2}    .4236{col 37}{space 2} .1339479{col 48}{space 1}    3.16{col 57}{space 3}0.002{col 65}{space 4} .1610669{col 78}{space 3} .6861331
{txt}{space 15}civtot_a {c |}{col 25}{res}{space 2}-.0531067{col 37}{space 2} .0936277{col 48}{space 1}   -0.57{col 57}{space 3}0.571{col 65}{space 4}-.2366137{col 78}{space 3} .1304002
{txt}{space 11}rival_weapon {c |}{col 25}{res}{space 2} 2.326273{col 37}{space 2}   .45773{col 48}{space 1}    5.08{col 57}{space 3}0.000{col 65}{space 4} 1.429139{col 78}{space 3} 3.223407
{txt}{space 17}gdp_pc {c |}{col 25}{res}{space 2}-.0241537{col 37}{space 2} .0291724{col 48}{space 1}   -0.83{col 57}{space 3}0.408{col 65}{space 4}-.0813305{col 78}{space 3} .0330231
{txt}{space 14}polity2_a {c |}{col 25}{res}{space 2}-.0470419{col 37}{space 2} .0431155{col 48}{space 1}   -1.09{col 57}{space 3}0.275{col 65}{space 4}-.1315467{col 78}{space 3} .0374628
{txt}{space 18}_cons {c |}{col 25}{res}{space 2} -5.74051{col 37}{space 2} .5692437{col 48}{space 1}  -10.08{col 57}{space 3}0.000{col 65}{space 4}-6.856207{col 78}{space 3}-4.624812
{txt}{hline 24}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *nu-clear Weighted (Column 3)
. logit audience_start tol_end_infw_total_full_5 Nat_infw_total_full_5 dll_infw_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-120.02512}  
Iteration 1:{space 3}log pseudolikelihood = {res:-106.10459}  
Iteration 2:{space 3}log pseudolikelihood = {res:-99.067165}  
Iteration 3:{space 3}log pseudolikelihood = {res:-98.371992}  
Iteration 4:{space 3}log pseudolikelihood = {res:-98.367329}  
Iteration 5:{space 3}log pseudolikelihood = {res:-98.367324}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:2,982}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:106.78}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-98.367324}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1804}

{txt}{ralign 91:(Std. err. adjusted for {res:100} clusters in {res:audience})}
{hline 26}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 27}{c |}{col 39}    Robust
{col 1}           audience_start{col 27}{c |} Coefficient{col 39}  std. err.{col 51}      z{col 59}   P>|z|{col 67}     [95% con{col 80}f. interval]
{hline 26}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_infw_total_full_5 {c |}{col 27}{res}{space 2} 1.501484{col 39}{space 2} .3072981{col 50}{space 1}    4.89{col 59}{space 3}0.000{col 67}{space 4} .8991905{col 80}{space 3} 2.103777
{txt}{space 4}Nat_infw_total_full_5 {c |}{col 27}{res}{space 2}-.2651547{col 39}{space 2} .2597015{col 50}{space 1}   -1.02{col 59}{space 3}0.307{col 67}{space 4}-.7741603{col 80}{space 3} .2438508
{txt}{space 4}dll_infw_total_full_5 {c |}{col 27}{res}{space 2} .1386918{col 39}{space 2} .5466943{col 50}{space 1}    0.25{col 59}{space 3}0.800{col 67}{space 4}-.9328093{col 80}{space 3} 1.210193
{txt}{space 17}inttot_a {c |}{col 27}{res}{space 2} .4147053{col 39}{space 2} .1276914{col 50}{space 1}    3.25{col 59}{space 3}0.001{col 67}{space 4} .1644347{col 80}{space 3} .6649759
{txt}{space 17}civtot_a {c |}{col 27}{res}{space 2}-.0651277{col 39}{space 2} .1147888{col 50}{space 1}   -0.57{col 59}{space 3}0.570{col 67}{space 4}-.2901095{col 80}{space 3} .1598542
{txt}{space 13}rival_weapon {c |}{col 27}{res}{space 2} 1.963364{col 39}{space 2} .5118244{col 50}{space 1}    3.84{col 59}{space 3}0.000{col 67}{space 4} .9602068{col 80}{space 3} 2.966522
{txt}{space 19}gdp_pc {c |}{col 27}{res}{space 2}-.0326546{col 39}{space 2} .0407789{col 50}{space 1}   -0.80{col 59}{space 3}0.423{col 67}{space 4}-.1125798{col 80}{space 3} .0472705
{txt}{space 16}polity2_a {c |}{col 27}{res}{space 2}-.0305461{col 39}{space 2} .0403779{col 50}{space 1}   -0.76{col 59}{space 3}0.449{col 67}{space 4}-.1096854{col 80}{space 3} .0485932
{txt}{space 20}_cons {c |}{col 27}{res}{space 2}-6.002773{col 39}{space 2} .5165987{col 50}{space 1}  -11.62{col 59}{space 3}0.000{col 67}{space 4}-7.015288{col 80}{space 3}-4.990259
{txt}{hline 26}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Ideal Point Weighted (Column 4)
. logit audience_start tol_end_ipdw_total_full_5 Nat_ipdw_total_full_5 dll_ipdw_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-49.926189}  
Iteration 1:{space 3}log pseudolikelihood = {res:-46.747235}  
Iteration 2:{space 3}log pseudolikelihood = {res:-39.307336}  
Iteration 3:{space 3}log pseudolikelihood = {res:-39.186394}  
Iteration 4:{space 3}log pseudolikelihood = {res:-39.184614}  
Iteration 5:{space 3}log pseudolikelihood = {res:-39.184613}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:3,227}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:119.09}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-39.184613}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.2151}

{txt}{ralign 91:(Std. err. adjusted for {res:119} clusters in {res:audience})}
{hline 26}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 27}{c |}{col 39}    Robust
{col 1}           audience_start{col 27}{c |} Coefficient{col 39}  std. err.{col 51}      z{col 59}   P>|z|{col 67}     [95% con{col 80}f. interval]
{hline 26}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_ipdw_total_full_5 {c |}{col 27}{res}{space 2} 1.261189{col 39}{space 2} .7594472{col 50}{space 1}    1.66{col 59}{space 3}0.097{col 67}{space 4}-.2273002{col 80}{space 3} 2.749678
{txt}{space 4}Nat_ipdw_total_full_5 {c |}{col 27}{res}{space 2}-.5127886{col 39}{space 2} .4028163{col 50}{space 1}   -1.27{col 59}{space 3}0.203{col 67}{space 4}-1.302294{col 80}{space 3} .2767168
{txt}{space 4}dll_ipdw_total_full_5 {c |}{col 27}{res}{space 2} .2023339{col 39}{space 2} .5769924{col 50}{space 1}    0.35{col 59}{space 3}0.726{col 67}{space 4}-.9285505{col 80}{space 3} 1.333218
{txt}{space 17}inttot_a {c |}{col 27}{res}{space 2} .6170135{col 39}{space 2} .2500946{col 50}{space 1}    2.47{col 59}{space 3}0.014{col 67}{space 4} .1268371{col 80}{space 3}  1.10719
{txt}{space 17}civtot_a {c |}{col 27}{res}{space 2}-.0872894{col 39}{space 2} .1797423{col 50}{space 1}   -0.49{col 59}{space 3}0.627{col 67}{space 4}-.4395779{col 80}{space 3}  .264999
{txt}{space 13}rival_weapon {c |}{col 27}{res}{space 2} 2.937125{col 39}{space 2} .6196218{col 50}{space 1}    4.74{col 59}{space 3}0.000{col 67}{space 4} 1.722688{col 80}{space 3} 4.151561
{txt}{space 19}gdp_pc {c |}{col 27}{res}{space 2}-.0029798{col 39}{space 2} .0207621{col 50}{space 1}   -0.14{col 59}{space 3}0.886{col 67}{space 4}-.0436727{col 80}{space 3} .0377131
{txt}{space 16}polity2_a {c |}{col 27}{res}{space 2}-.0935204{col 39}{space 2} .0755106{col 50}{space 1}   -1.24{col 59}{space 3}0.216{col 67}{space 4}-.2415185{col 80}{space 3} .0544778
{txt}{space 20}_cons {c |}{col 27}{res}{space 2}-6.776351{col 39}{space 2} .7715022{col 50}{space 1}   -8.78{col 59}{space 3}0.000{col 67}{space 4}-8.288467{col 80}{space 3}-5.264234
{txt}{hline 26}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Polity Weighted (Column 5)
. logit audience_start tol_end_pw_total_full_5 Nat_pw_total_full_5 dll_pw_total_full_5 inttot_a civtot_a rival_weapon gdp_pc polity2_a  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-120.34337}  
Iteration 1:{space 3}log pseudolikelihood = {res:-107.76865}  
Iteration 2:{space 3}log pseudolikelihood = {res:-101.32257}  
Iteration 3:{space 3}log pseudolikelihood = {res:-100.49382}  
Iteration 4:{space 3}log pseudolikelihood = {res:-100.48085}  
Iteration 5:{space 3}log pseudolikelihood = {res:-100.48082}  
Iteration 6:{space 3}log pseudolikelihood = {res:-100.48082}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:3,947}
{txt}{col 57}{lalign 13:Wald chi2({res:8})}{col 70} = {res}{ralign 6:133.44}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-100.48082}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1650}

{txt}{ralign 89:(Std. err. adjusted for {res:120} clusters in {res:audience})}
{hline 24}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 25}{c |}{col 37}    Robust
{col 1}         audience_start{col 25}{c |} Coefficient{col 37}  std. err.{col 49}      z{col 57}   P>|z|{col 65}     [95% con{col 78}f. interval]
{hline 24}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_pw_total_full_5 {c |}{col 25}{res}{space 2} 1.681577{col 37}{space 2}  .434965{col 48}{space 1}    3.87{col 57}{space 3}0.000{col 65}{space 4} .8290613{col 78}{space 3} 2.534093
{txt}{space 4}Nat_pw_total_full_5 {c |}{col 25}{res}{space 2} -.468756{col 37}{space 2} .2699997{col 48}{space 1}   -1.74{col 57}{space 3}0.083{col 65}{space 4}-.9979457{col 78}{space 3} .0604337
{txt}{space 4}dll_pw_total_full_5 {c |}{col 25}{res}{space 2} .5650453{col 37}{space 2} .6142396{col 48}{space 1}    0.92{col 57}{space 3}0.358{col 65}{space 4}-.6388422{col 78}{space 3} 1.768933
{txt}{space 15}inttot_a {c |}{col 25}{res}{space 2} .2756864{col 37}{space 2} .1675744{col 48}{space 1}    1.65{col 57}{space 3}0.100{col 65}{space 4}-.0527534{col 78}{space 3} .6041261
{txt}{space 15}civtot_a {c |}{col 25}{res}{space 2}-.0437388{col 37}{space 2} .1029509{col 48}{space 1}   -0.42{col 57}{space 3}0.671{col 65}{space 4}-.2455189{col 78}{space 3} .1580414
{txt}{space 11}rival_weapon {c |}{col 25}{res}{space 2} 2.115421{col 37}{space 2} .5287284{col 48}{space 1}    4.00{col 57}{space 3}0.000{col 65}{space 4} 1.079132{col 78}{space 3} 3.151709
{txt}{space 17}gdp_pc {c |}{col 25}{res}{space 2}-.0265313{col 37}{space 2} .0298568{col 48}{space 1}   -0.89{col 57}{space 3}0.374{col 65}{space 4}-.0850496{col 78}{space 3} .0319869
{txt}{space 14}polity2_a {c |}{col 25}{res}{space 2}-.0694919{col 37}{space 2}  .042517{col 48}{space 1}   -1.63{col 57}{space 3}0.102{col 65}{space 4}-.1528236{col 78}{space 3} .0138398
{txt}{space 18}_cons {c |}{col 25}{res}{space 2}-6.096636{col 37}{space 2} .5343848{col 48}{space 1}  -11.41{col 57}{space 3}0.000{col 65}{space 4}-7.144011{col 78}{space 3}-5.049261
{txt}{hline 24}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. 
. 
. *** APPENDIX TABLE 8 ***
. 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-204.09675}  
Iteration 1:{space 3}log pseudolikelihood = {res:-197.14631}  
Iteration 2:{space 3}log pseudolikelihood = {res:-195.50532}  
Iteration 3:{space 3}log pseudolikelihood = {res:-195.49468}  
Iteration 4:{space 3}log pseudolikelihood = {res:-195.49467}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:9,956}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:18.32}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0004}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-195.49467}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0421}

{txt}{ralign 86:(Std. err. adjusted for {res:197} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .5516382{col 34}{space 2} .2363715{col 45}{space 1}    2.33{col 54}{space 3}0.020{col 62}{space 4} .0883586{col 75}{space 3} 1.014918
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.4973543{col 34}{space 2} .2205432{col 45}{space 1}   -2.26{col 54}{space 3}0.024{col 62}{space 4}-.9296109{col 75}{space 3}-.0650976
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.2628326{col 34}{space 2} .2605141{col 45}{space 1}   -1.01{col 54}{space 3}0.313{col 62}{space 4}-.7734308{col 75}{space 3} .2477655
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.747665{col 34}{space 2} .4001985{col 45}{space 1}  -14.36{col 54}{space 3}0.000{col 62}{space 4} -6.53204{col 75}{space 3} -4.96329
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5   if (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-132.14233}  
Iteration 1:{space 3}log pseudolikelihood = {res:-126.77483}  
Iteration 2:{space 3}log pseudolikelihood = {res:-124.58613}  
Iteration 3:{space 3}log pseudolikelihood = {res:-124.47007}  
Iteration 4:{space 3}log pseudolikelihood = {res:-124.46944}  
Iteration 5:{space 3}log pseudolikelihood = {res:-124.46944}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:10,224}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:7.82}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0498}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-124.46944}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0581}

{txt}{ralign 86:(Std. err. adjusted for {res:198} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .5119041{col 34}{space 2} .3654706{col 45}{space 1}    1.40{col 54}{space 3}0.161{col 62}{space 4}-.2044051{col 75}{space 3} 1.228213
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.266846{col 34}{space 2} .5004785{col 45}{space 1}   -2.53{col 54}{space 3}0.011{col 62}{space 4}-2.247766{col 75}{space 3}-.2859261
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .2667318{col 34}{space 2} .3459594{col 45}{space 1}    0.77{col 54}{space 3}0.441{col 62}{space 4}-.4113362{col 75}{space 3} .9447997
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.321577{col 34}{space 2} .5860649{col 45}{space 1}  -10.79{col 54}{space 3}0.000{col 62}{space 4}-7.470243{col 75}{space 3} -5.17291
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-186.52774}  
Iteration 1:{space 3}log pseudolikelihood = {res: -178.7888}  
Iteration 2:{space 3}log pseudolikelihood = {res:-175.59232}  
Iteration 3:{space 3}log pseudolikelihood = {res:-175.56283}  
Iteration 4:{space 3}log pseudolikelihood = {res:-175.56281}  
Iteration 5:{space 3}log pseudolikelihood = {res:-175.56281}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:9,953}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:27.79}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-175.56281}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0588}

{txt}{ralign 86:(Std. err. adjusted for {res:197} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .7521615{col 34}{space 2} .2518963{col 45}{space 1}    2.99{col 54}{space 3}0.003{col 62}{space 4} .2584539{col 75}{space 3} 1.245869
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.4732894{col 34}{space 2} .2391594{col 45}{space 1}   -1.98{col 54}{space 3}0.048{col 62}{space 4}-.9420332{col 75}{space 3}-.0045456
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.4044857{col 34}{space 2} .3305539{col 45}{space 1}   -1.22{col 54}{space 3}0.221{col 62}{space 4}-1.052359{col 75}{space 3}  .243388
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.030897{col 34}{space 2} .4424941{col 45}{space 1}  -13.63{col 54}{space 3}0.000{col 62}{space 4}-6.898169{col 75}{space 3}-5.163624
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if year_a>1945 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-283.63054}  
Iteration 1:{space 3}log pseudolikelihood = {res:-272.18464}  
Iteration 2:{space 3}log pseudolikelihood = {res:-268.99567}  
Iteration 3:{space 3}log pseudolikelihood = {res:-268.93661}  
Iteration 4:{space 3}log pseudolikelihood = {res:-268.93657}  
Iteration 5:{space 3}log pseudolikelihood = {res:-268.93657}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:10,224}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:29.37}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-268.93657}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0518}

{txt}{ralign 86:(Std. err. adjusted for {res:198} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}  .585242{col 34}{space 2} .2240022{col 45}{space 1}    2.61{col 54}{space 3}0.009{col 62}{space 4} .1462057{col 75}{space 3} 1.024278
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.6739899{col 34}{space 2} .2356779{col 45}{space 1}   -2.86{col 54}{space 3}0.004{col 62}{space 4} -1.13591{col 75}{space 3}-.2120696
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.1279775{col 34}{space 2}  .212496{col 45}{space 1}   -0.60{col 54}{space 3}0.547{col 62}{space 4} -.544462{col 75}{space 3}  .288507
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.437838{col 34}{space 2} .3653127{col 45}{space 1}  -14.89{col 54}{space 3}0.000{col 62}{space 4}-6.153838{col 75}{space 3}-4.721838
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if  resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-97.941224}  
Iteration 1:{space 3}log pseudolikelihood = {res:-91.890155}  
Iteration 2:{space 3}log pseudolikelihood = {res:-89.851029}  
Iteration 3:{space 3}log pseudolikelihood = {res:-89.844179}  
Iteration 4:{space 3}log pseudolikelihood = {res:-89.844178}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:26.69}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-89.844178}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0827}

{txt}{ralign 86:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-1.028042{col 34}{space 2} .2722958{col 45}{space 1}   -3.78{col 54}{space 3}0.000{col 62}{space 4}-1.561732{col 75}{space 3} -.494352
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3181166{col 34}{space 2} .1188786{col 45}{space 1}    2.68{col 54}{space 3}0.007{col 62}{space 4} .0851189{col 75}{space 3} .5511143
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.9699223{col 34}{space 2} .4257929{col 45}{space 1}   -2.28{col 54}{space 3}0.023{col 62}{space 4}-1.804461{col 75}{space 3}-.1353836
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-1.930569{col 34}{space 2} .3243576{col 45}{space 1}   -5.95{col 54}{space 3}0.000{col 62}{space 4}-2.566299{col 75}{space 3} -1.29484
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if   resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res: -95.02947}  
Iteration 1:{space 3}log pseudolikelihood = {res:-89.268655}  
Iteration 2:{space 3}log pseudolikelihood = {res:-86.851685}  
Iteration 3:{space 3}log pseudolikelihood = {res:-86.842892}  
Iteration 4:{space 3}log pseudolikelihood = {res: -86.84289}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:25.28}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-86.84289}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0861}

{txt}{ralign 86:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-1.065486{col 34}{space 2} .2921816{col 45}{space 1}   -3.65{col 54}{space 3}0.000{col 62}{space 4}-1.638151{col 75}{space 3}-.4928206
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3400882{col 34}{space 2} .1214351{col 45}{space 1}    2.80{col 54}{space 3}0.005{col 62}{space 4} .1020797{col 75}{space 3} .5780967
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -.935321{col 34}{space 2} .4292657{col 45}{space 1}   -2.18{col 54}{space 3}0.029{col 62}{space 4}-1.776666{col 75}{space 3}-.0939756
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -1.97955{col 34}{space 2} .3414956{col 45}{space 1}   -5.80{col 54}{space 3}0.000{col 62}{space 4}-2.648869{col 75}{space 3} -1.31023
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if  year_a>1945, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}    10,423
                                                {txt}F(3, 199)         =  {res}     4.38
                                                {txt}Prob > F          = {res}    0.0052
                                                {txt}R-squared         = {res}    0.0016
                                                {txt}Root MSE          =    {res} .08247

{txt}{ralign 86:(Std. err. adjusted for {res:200} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0037642{col 34}{space 2} .0016213{col 45}{space 1}    2.32{col 54}{space 3}0.021{col 62}{space 4} .0005671{col 75}{space 3} .0069613
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0015599{col 34}{space 2} .0007037{col 45}{space 1}   -2.22{col 54}{space 3}0.028{col 62}{space 4}-.0029476{col 75}{space 3}-.0001723
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0009291{col 34}{space 2} .0008705{col 45}{space 1}    1.07{col 54}{space 3}0.287{col 62}{space 4}-.0007876{col 75}{space 3} .0026457
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-2.52e-07{col 34}{space 2} .0019073{col 45}{space 1}   -0.00{col 54}{space 3}1.000{col 62}{space 4}-.0037613{col 75}{space 3} .0037608
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. *** APPENDIX TABLE 9***
. 
.                 
. *Program (Column 1)
. logit audience_start  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-185.77755}  
Iteration 1:{space 3}log pseudolikelihood = {res: -177.0528}  
Iteration 2:{space 3}log pseudolikelihood = {res:-175.22358}  
Iteration 3:{space 3}log pseudolikelihood = {res:-175.20013}  
Iteration 4:{space 3}log pseudolikelihood = {res:-175.20007}  
Iteration 5:{space 3}log pseudolikelihood = {res:-175.20007}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,413}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:17.69}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0005}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-175.20007}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0569}

{txt}{ralign 92:(Std. err. adjusted for {res:140} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}            audience_start{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2} 1.636241{col 40}{space 2} .4677174{col 51}{space 1}    3.50{col 60}{space 3}0.000{col 68}{space 4} .7195315{col 81}{space 3}  2.55295
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2}-1.004151{col 40}{space 2} .3561918{col 51}{space 1}   -2.82{col 60}{space 3}0.005{col 68}{space 4}-1.702274{col 81}{space 3}-.3060276
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2} .1234036{col 40}{space 2} .4209925{col 51}{space 1}    0.29{col 60}{space 3}0.769{col 68}{space 4}-.7017265{col 81}{space 3} .9485338
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-5.967874{col 40}{space 2} .5026271{col 51}{space 1}  -11.87{col 60}{space 3}0.000{col 68}{space 4}-6.953005{col 81}{space 3}-4.982743
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-121.55265}  
Iteration 1:{space 3}log pseudolikelihood = {res:-115.05502}  
Iteration 2:{space 3}log pseudolikelihood = {res:-112.91055}  
Iteration 3:{space 3}log pseudolikelihood = {res:-112.88835}  
Iteration 4:{space 3}log pseudolikelihood = {res:-112.88833}  
Iteration 5:{space 3}log pseudolikelihood = {res:-112.88833}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,681}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:11.46}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0095}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-112.88833}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0713}

{txt}{ralign 92:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}              pursue_start{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2} 1.430757{col 40}{space 2} .5984686{col 51}{space 1}    2.39{col 60}{space 3}0.017{col 68}{space 4} .2577804{col 81}{space 3} 2.603734
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2}-1.626896{col 40}{space 2} .6115161{col 51}{space 1}   -2.66{col 60}{space 3}0.008{col 68}{space 4}-2.825445{col 81}{space 3}-.4283462
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2} .8543419{col 40}{space 2} .4294306{col 51}{space 1}    1.99{col 60}{space 3}0.047{col 68}{space 4} .0126735{col 81}{space 3}  1.69601
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-6.573459{col 40}{space 2} .6640506{col 51}{space 1}   -9.90{col 60}{space 3}0.000{col 68}{space 4}-7.874975{col 81}{space 3}-5.271944
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-170.03704}  
Iteration 1:{space 3}log pseudolikelihood = {res:-159.69916}  
Iteration 2:{space 3}log pseudolikelihood = {res:-156.36899}  
Iteration 3:{space 3}log pseudolikelihood = {res:-156.28576}  
Iteration 4:{space 3}log pseudolikelihood = {res:-156.28538}  
Iteration 5:{space 3}log pseudolikelihood = {res:-156.28538}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,410}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:23.21}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-156.28538}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0809}

{txt}{ralign 92:(Std. err. adjusted for {res:140} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}             explore_start{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2} 2.181485{col 40}{space 2} .7431957{col 51}{space 1}    2.94{col 60}{space 3}0.003{col 68}{space 4} .7248477{col 81}{space 3} 3.638121
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2}-1.089806{col 40}{space 2} .3702943{col 51}{space 1}   -2.94{col 60}{space 3}0.003{col 68}{space 4}-1.815569{col 81}{space 3}-.3640424
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2}-.2636158{col 40}{space 2} .4815907{col 51}{space 1}   -0.55{col 60}{space 3}0.584{col 68}{space 4}-1.207516{col 81}{space 3} .6802847
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-6.402036{col 40}{space 2}  .773788{col 51}{space 1}   -8.27{col 60}{space 3}0.000{col 68}{space 4}-7.918632{col 81}{space 3}-4.885439
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & year_a>1945 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-257.69965}  
Iteration 1:{space 3}log pseudolikelihood = {res: -243.0959}  
Iteration 2:{space 3}log pseudolikelihood = {res:-239.50685}  
Iteration 3:{space 3}log pseudolikelihood = {res:-239.44229}  
Iteration 4:{space 3}log pseudolikelihood = {res:-239.44224}  
Iteration 5:{space 3}log pseudolikelihood = {res:-239.44224}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,681}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:23.57}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-239.44224}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0708}

{txt}{ralign 92:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}            audience_accel{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2} 1.756344{col 40}{space 2} .5350967{col 51}{space 1}    3.28{col 60}{space 3}0.001{col 68}{space 4}  .707574{col 81}{space 3} 2.805115
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2}-1.228216{col 40}{space 2} .3304428{col 51}{space 1}   -3.72{col 60}{space 3}0.000{col 68}{space 4}-1.875872{col 81}{space 3}-.5805601
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2} .2436722{col 40}{space 2} .3120074{col 51}{space 1}    0.78{col 60}{space 3}0.435{col 68}{space 4}-.3678512{col 81}{space 3} .8551955
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-5.755986{col 40}{space 2} .5553797{col 51}{space 1}  -10.36{col 60}{space 3}0.000{col 68}{space 4} -6.84451{col 81}{space 3}-4.667462
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-97.941224}  
Iteration 1:{space 3}log pseudolikelihood = {res:-93.865523}  
Iteration 2:{space 3}log pseudolikelihood = {res:-93.451963}  
Iteration 3:{space 3}log pseudolikelihood = {res:-93.451349}  
Iteration 4:{space 3}log pseudolikelihood = {res:-93.451349}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:12.24}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0066}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-93.451349}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0458}

{txt}{ralign 92:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}            audience_decel{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2}-.9631618{col 40}{space 2}  .372356{col 51}{space 1}   -2.59{col 60}{space 3}0.010{col 68}{space 4}-1.692966{col 81}{space 3}-.2333574
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2} .7370397{col 40}{space 2} .3476144{col 51}{space 1}    2.12{col 60}{space 3}0.034{col 68}{space 4}  .055728{col 81}{space 3} 1.418351
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2}-1.041998{col 40}{space 2} .6375039{col 51}{space 1}   -1.63{col 60}{space 3}0.102{col 68}{space 4}-2.291482{col 81}{space 3} .2074871
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-2.186279{col 40}{space 2} .3827762{col 51}{space 1}   -5.71{col 60}{space 3}0.000{col 68}{space 4}-2.936507{col 81}{space 3}-1.436052
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res: -95.02947}  
Iteration 1:{space 3}log pseudolikelihood = {res:-90.758177}  
Iteration 2:{space 3}log pseudolikelihood = {res:-90.248734}  
Iteration 3:{space 3}log pseudolikelihood = {res: -90.24809}  
Iteration 4:{space 3}log pseudolikelihood = {res: -90.24809}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:475}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:13.19}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0042}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-90.24809}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0503}

{txt}{ralign 92:(Std. err. adjusted for {res:30} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}              audience_end{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      z{col 60}   P>|z|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2}-1.025858{col 40}{space 2} .3840616{col 51}{space 1}   -2.67{col 60}{space 3}0.008{col 68}{space 4}-1.778605{col 81}{space 3}-.2731113
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2} .8167329{col 40}{space 2} .3668047{col 51}{space 1}    2.23{col 60}{space 3}0.026{col 68}{space 4}  .097809{col 81}{space 3} 1.535657
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2}-1.023438{col 40}{space 2} .6522086{col 51}{space 1}   -1.57{col 60}{space 3}0.117{col 68}{space 4}-2.301744{col 81}{space 3} .2548672
{txt}{space 21}_cons {c |}{col 28}{res}{space 2}-2.232698{col 40}{space 2} .3982128{col 51}{space 1}   -5.61{col 60}{space 3}0.000{col 68}{space 4}-3.013181{col 81}{space 3}-1.452215
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  dummy_tol_end_total_full_5 dummy_at_total_full_5 dummy_dll_total_full_5  if aec2==1 & year_a>1945, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     5,880
                                                {txt}F(3, 142)         =  {res}     4.43
                                                {txt}Prob > F          = {res}    0.0052
                                                {txt}R-squared         = {res}    0.0030
                                                {txt}Root MSE          =    {res} .10972

{txt}{ralign 92:(Std. err. adjusted for {res:143} clusters in {res:audience})}
{hline 27}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 28}{c |}{col 40}    Robust
{col 1}             status_change{col 28}{c |} Coefficient{col 40}  std. err.{col 52}      t{col 60}   P>|t|{col 68}     [95% con{col 81}f. interval]
{hline 27}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
dummy_tol_end_total_full_5 {c |}{col 28}{res}{space 2} .0085464{col 40}{space 2} .0034135{col 51}{space 1}    2.50{col 60}{space 3}0.013{col 68}{space 4} .0017986{col 81}{space 3} .0152941
{txt}{space 5}dummy_at_total_full_5 {c |}{col 28}{res}{space 2}-.0085247{col 40}{space 2}  .003082{col 51}{space 1}   -2.77{col 60}{space 3}0.006{col 68}{space 4}-.0146173{col 81}{space 3}-.0024321
{txt}{space 4}dummy_dll_total_full_5 {c |}{col 28}{res}{space 2} .0057279{col 40}{space 2} .0032831{col 51}{space 1}    1.74{col 60}{space 3}0.083{col 68}{space 4}-.0007621{col 81}{space 3} .0122179
{txt}{space 21}_cons {c |}{col 28}{res}{space 2} .0001996{col 40}{space 2} .0032812{col 51}{space 1}    0.06{col 60}{space 3}0.952{col 68}{space 4}-.0062866{col 81}{space 3} .0066859
{txt}{hline 27}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. 
. 
. *** APPENDIX TABLE 10***
.                 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>=1939, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-226.64392}  
Iteration 1:{space 3}log pseudolikelihood = {res:  -216.689}  
Iteration 2:{space 3}log pseudolikelihood = {res:-214.41427}  
Iteration 3:{space 3}log pseudolikelihood = {res:-214.40407}  
Iteration 4:{space 3}log pseudolikelihood = {res:-214.40407}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,461}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:28.02}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-214.40407}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0540}

{txt}{ralign 86:(Std. err. adjusted for {res:144} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .3164284{col 34}{space 2} .2142953{col 45}{space 1}    1.48{col 54}{space 3}0.140{col 62}{space 4}-.1035826{col 75}{space 3} .7364394
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} -.530638{col 34}{space 2} .1642429{col 45}{space 1}   -3.23{col 54}{space 3}0.001{col 62}{space 4}-.8525482{col 75}{space 3}-.2087277
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.5826528{col 34}{space 2} .2635684{col 45}{space 1}   -2.21{col 54}{space 3}0.027{col 62}{space 4}-1.099237{col 75}{space 3}-.0660682
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-4.461699{col 34}{space 2} .3271793{col 45}{space 1}  -13.64{col 54}{space 3}0.000{col 62}{space 4}-5.102959{col 75}{space 3}-3.820439
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>=1939, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-138.80073}  
Iteration 1:{space 3}log pseudolikelihood = {res:-132.76648}  
Iteration 2:{space 3}log pseudolikelihood = {res:-130.85647}  
Iteration 3:{space 3}log pseudolikelihood = {res:-130.75015}  
Iteration 4:{space 3}log pseudolikelihood = {res:-130.74961}  
Iteration 5:{space 3}log pseudolikelihood = {res:-130.74961}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,744}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:11.43}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0096}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-130.74961}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0580}

{txt}{ralign 86:(Std. err. adjusted for {res:144} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .2088916{col 34}{space 2} .3576387{col 45}{space 1}    0.58{col 54}{space 3}0.559{col 62}{space 4}-.4920674{col 75}{space 3} .9098506
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.113152{col 34}{space 2} .3628876{col 45}{space 1}   -3.07{col 54}{space 3}0.002{col 62}{space 4}-1.824399{col 75}{space 3}-.4019056
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.0451226{col 34}{space 2} .3178009{col 45}{space 1}   -0.14{col 54}{space 3}0.887{col 62}{space 4}-.6680009{col 75}{space 3} .5777558
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -5.11131{col 34}{space 2} .4900305{col 45}{space 1}  -10.43{col 54}{space 3}0.000{col 62}{space 4}-6.071752{col 75}{space 3}-4.150868
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>=1939, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-211.61968}  
Iteration 1:{space 3}log pseudolikelihood = {res:-200.79245}  
Iteration 2:{space 3}log pseudolikelihood = {res:-197.11856}  
Iteration 3:{space 3}log pseudolikelihood = {res:-197.08003}  
Iteration 4:{space 3}log pseudolikelihood = {res:-197.07998}  
Iteration 5:{space 3}log pseudolikelihood = {res:-197.07998}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,458}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:35.63}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-197.07998}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0687}

{txt}{ralign 86:(Std. err. adjusted for {res:144} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}  .440545{col 34}{space 2} .2350718{col 45}{space 1}    1.87{col 54}{space 3}0.061{col 62}{space 4}-.0201874{col 75}{space 3} .9012773
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5175032{col 34}{space 2} .1726948{col 45}{space 1}   -3.00{col 54}{space 3}0.003{col 62}{space 4}-.8559788{col 75}{space 3}-.1790276
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.7691125{col 34}{space 2} .3261928{col 45}{space 1}   -2.36{col 54}{space 3}0.018{col 62}{space 4}-1.408439{col 75}{space 3}-.1297863
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-4.608812{col 34}{space 2}  .359384{col 45}{space 1}  -12.82{col 54}{space 3}0.000{col 62}{space 4}-5.313192{col 75}{space 3}-3.904432
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1939 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-310.40762}  
Iteration 1:{space 3}log pseudolikelihood = {res:-295.80984}  
Iteration 2:{space 3}log pseudolikelihood = {res:-292.69242}  
Iteration 3:{space 3}log pseudolikelihood = {res:  -292.649}  
Iteration 4:{space 3}log pseudolikelihood = {res:-292.64886}  
Iteration 5:{space 3}log pseudolikelihood = {res:-292.64886}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:5,744}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:42.37}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-292.64886}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0572}

{txt}{ralign 86:(Std. err. adjusted for {res:144} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .2754325{col 34}{space 2}  .235595{col 45}{space 1}    1.17{col 54}{space 3}0.242{col 62}{space 4}-.1863252{col 75}{space 3} .7371901
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.6505117{col 34}{space 2} .1662831{col 45}{space 1}   -3.91{col 54}{space 3}0.000{col 62}{space 4}-.9764205{col 75}{space 3}-.3246029
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -.460719{col 34}{space 2} .2246896{col 45}{space 1}   -2.05{col 54}{space 3}0.040{col 62}{space 4}-.9011026{col 75}{space 3}-.0203354
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -4.13559{col 34}{space 2} .3356416{col 45}{space 1}  -12.32{col 54}{space 3}0.000{col 62}{space 4}-4.793435{col 75}{space 3}-3.477744
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>=1939, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-99.308862}  
Iteration 1:{space 3}log pseudolikelihood = {res:-94.655458}  
Iteration 2:{space 3}log pseudolikelihood = {res:-93.016067}  
Iteration 3:{space 3}log pseudolikelihood = {res:-93.011452}  
Iteration 4:{space 3}log pseudolikelihood = {res:-93.011451}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:501}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:20.20}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0002}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-93.011451}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0634}

{txt}{ralign 86:(Std. err. adjusted for {res:32} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.7892417{col 34}{space 2} .2755394{col 45}{space 1}   -2.86{col 54}{space 3}0.004{col 62}{space 4}-1.329289{col 75}{space 3}-.2491944
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3313484{col 34}{space 2} .1251637{col 45}{space 1}    2.65{col 54}{space 3}0.008{col 62}{space 4} .0860321{col 75}{space 3} .5766648
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.7897949{col 34}{space 2} .4187721{col 45}{space 1}   -1.89{col 54}{space 3}0.059{col 62}{space 4}-1.610573{col 75}{space 3} .0309833
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -2.28362{col 34}{space 2} .3497094{col 45}{space 1}   -6.53{col 54}{space 3}0.000{col 62}{space 4}-2.969038{col 75}{space 3}-1.598202
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>=1939, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-96.340999}  
Iteration 1:{space 3}log pseudolikelihood = {res:-91.910561}  
Iteration 2:{space 3}log pseudolikelihood = {res:-89.902218}  
Iteration 3:{space 3}log pseudolikelihood = {res:-89.895598}  
Iteration 4:{space 3}log pseudolikelihood = {res:-89.895596}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:501}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:19.80}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0002}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-89.895596}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0669}

{txt}{ralign 86:(Std. err. adjusted for {res:32} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.8259405{col 34}{space 2} .2937953{col 45}{space 1}   -2.81{col 54}{space 3}0.005{col 62}{space 4}-1.401769{col 75}{space 3}-.2501124
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3542657{col 34}{space 2} .1274824{col 45}{space 1}    2.78{col 54}{space 3}0.005{col 62}{space 4} .1044048{col 75}{space 3} .6041267
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.7535673{col 34}{space 2} .4214578{col 45}{space 1}   -1.79{col 54}{space 3}0.074{col 62}{space 4}-1.579609{col 75}{space 3} .0724747
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}   -2.334{col 34}{space 2} .3657888{col 45}{space 1}   -6.38{col 54}{space 3}0.000{col 62}{space 4}-3.050932{col 75}{space 3}-1.617067
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1939, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     5,952
                                                {txt}F(3, 143)         =  {res}     5.62
                                                {txt}Prob > F          = {res}    0.0011
                                                {txt}R-squared         = {res}    0.0025
                                                {txt}Root MSE          =    {res} .11861

{txt}{ralign 86:(Std. err. adjusted for {res:144} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0043816{col 34}{space 2} .0032086{col 45}{space 1}    1.37{col 54}{space 3}0.174{col 62}{space 4}-.0019608{col 75}{space 3} .0107239
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0037094{col 34}{space 2} .0012901{col 45}{space 1}   -2.88{col 54}{space 3}0.005{col 62}{space 4}-.0062595{col 75}{space 3}-.0011593
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.0008353{col 34}{space 2} .0018068{col 45}{space 1}   -0.46{col 54}{space 3}0.645{col 62}{space 4}-.0044068{col 75}{space 3} .0027363
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} .0062548{col 34}{space 2} .0042226{col 45}{space 1}    1.48{col 54}{space 3}0.141{col 62}{space 4}-.0020921{col 75}{space 3} .0146016
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. *** APPENDIX TABLE 11***
.                 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>=1969, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-106.16734}  
Iteration 1:{space 3}log pseudolikelihood = {res:-101.54919}  
Iteration 2:{space 3}log pseudolikelihood = {res:-99.446868}  
Iteration 3:{space 3}log pseudolikelihood = {res:-99.430182}  
Iteration 4:{space 3}log pseudolikelihood = {res:-99.430171}  
Iteration 5:{space 3}log pseudolikelihood = {res:-99.430171}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,491}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:10.10}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0178}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-99.430171}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0635}

{txt}{ralign 86:(Std. err. adjusted for {res:137} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} 1.002329{col 34}{space 2} .3657805{col 45}{space 1}    2.74{col 54}{space 3}0.006{col 62}{space 4} .2854125{col 75}{space 3} 1.719246
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.6574231{col 34}{space 2} .3473414{col 45}{space 1}   -1.89{col 54}{space 3}0.058{col 62}{space 4}  -1.3382{col 75}{space 3} .0233535
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.1039472{col 34}{space 2} .2688349{col 45}{space 1}   -0.39{col 54}{space 3}0.699{col 62}{space 4} -.630854{col 75}{space 3} .4229596
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.585744{col 34}{space 2} .4640634{col 45}{space 1}  -12.04{col 54}{space 3}0.000{col 62}{space 4}-6.495292{col 75}{space 3}-4.676197
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>=1969, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-89.322025}  
Iteration 1:{space 3}log pseudolikelihood = {res:-84.239212}  
Iteration 2:{space 3}log pseudolikelihood = {res:-81.833774}  
Iteration 3:{space 3}log pseudolikelihood = {res:-81.685448}  
Iteration 4:{space 3}log pseudolikelihood = {res:-81.684206}  
Iteration 5:{space 3}log pseudolikelihood = {res:-81.684206}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,616}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:13.95}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0030}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-81.684206}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0855}

{txt}{ralign 86:(Std. err. adjusted for {res:138} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .6827198{col 34}{space 2} .2928516{col 45}{space 1}    2.33{col 54}{space 3}0.020{col 62}{space 4} .1087413{col 75}{space 3} 1.256698
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.286928{col 34}{space 2} .4528873{col 45}{space 1}   -2.84{col 54}{space 3}0.004{col 62}{space 4}-2.174571{col 75}{space 3} -.399285
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0344532{col 34}{space 2} .3500859{col 45}{space 1}    0.10{col 54}{space 3}0.922{col 62}{space 4}-.6517026{col 75}{space 3}  .720609
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.441062{col 34}{space 2} .6013351{col 45}{space 1}   -9.05{col 54}{space 3}0.000{col 62}{space 4}-6.619657{col 75}{space 3}-4.262467
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>=1969, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-83.072335}  
Iteration 1:{space 3}log pseudolikelihood = {res:-79.827459}  
Iteration 2:{space 3}log pseudolikelihood = {res:-76.536333}  
Iteration 3:{space 3}log pseudolikelihood = {res:-76.525643}  
Iteration 4:{space 3}log pseudolikelihood = {res:-76.525636}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,487}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:12.08}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0071}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-76.525636}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0788}

{txt}{ralign 86:(Std. err. adjusted for {res:136} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} 1.321278{col 34}{space 2} .4300428{col 45}{space 1}    3.07{col 54}{space 3}0.002{col 62}{space 4} .4784097{col 75}{space 3} 2.164146
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.5868104{col 34}{space 2} .4039489{col 45}{space 1}   -1.45{col 54}{space 3}0.146{col 62}{space 4}-1.378536{col 75}{space 3} .2049149
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.1629922{col 34}{space 2} .3781845{col 45}{space 1}   -0.43{col 54}{space 3}0.666{col 62}{space 4}-.9042201{col 75}{space 3} .5782357
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.167664{col 34}{space 2} .5841244{col 45}{space 1}  -10.56{col 54}{space 3}0.000{col 62}{space 4}-7.312526{col 75}{space 3}-5.022801
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1969 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-155.39238}  
Iteration 1:{space 3}log pseudolikelihood = {res:-147.07656}  
Iteration 2:{space 3}log pseudolikelihood = {res:-142.97494}  
Iteration 3:{space 3}log pseudolikelihood = {res:-142.91523}  
Iteration 4:{space 3}log pseudolikelihood = {res:-142.91509}  
Iteration 5:{space 3}log pseudolikelihood = {res:-142.91509}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,616}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:19.22}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0002}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-142.91509}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0803}

{txt}{ralign 86:(Std. err. adjusted for {res:138} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .9517134{col 34}{space 2} .2928369{col 45}{space 1}    3.25{col 54}{space 3}0.001{col 62}{space 4} .3777637{col 75}{space 3} 1.525663
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.8560414{col 34}{space 2}  .340079{col 45}{space 1}   -2.52{col 54}{space 3}0.012{col 62}{space 4}-1.522584{col 75}{space 3}-.1894988
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.0743466{col 34}{space 2} .2047021{col 45}{space 1}   -0.36{col 54}{space 3}0.716{col 62}{space 4}-.4755553{col 75}{space 3} .3268621
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.071465{col 34}{space 2} .4216218{col 45}{space 1}  -12.03{col 54}{space 3}0.000{col 62}{space 4}-5.897828{col 75}{space 3}-4.245101
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>=1969, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-69.030557}  
Iteration 1:{space 3}log pseudolikelihood = {res:-63.541832}  
Iteration 2:{space 3}log pseudolikelihood = {res:-62.354694}  
Iteration 3:{space 3}log pseudolikelihood = {res:-62.348999}  
Iteration 4:{space 3}log pseudolikelihood = {res:-62.348996}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:274}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:15.21}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0016}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-62.348996}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0968}

{txt}{ralign 86:(Std. err. adjusted for {res:21} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.4914074{col 34}{space 2} .4595187{col 45}{space 1}   -1.07{col 54}{space 3}0.285{col 62}{space 4}-1.392048{col 75}{space 3} .4092327
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}  .171314{col 34}{space 2} .1311172{col 45}{space 1}    1.31{col 54}{space 3}0.191{col 62}{space 4} -.085671{col 75}{space 3}  .428299
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -1.28216{col 34}{space 2} .3605393{col 45}{space 1}   -3.56{col 54}{space 3}0.000{col 62}{space 4}-1.988804{col 75}{space 3}-.5755163
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -1.70061{col 34}{space 2} .3237881{col 45}{space 1}   -5.25{col 54}{space 3}0.000{col 62}{space 4}-2.335223{col 75}{space 3}-1.065997
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>=1969, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-69.030557}  
Iteration 1:{space 3}log pseudolikelihood = {res:-63.541832}  
Iteration 2:{space 3}log pseudolikelihood = {res:-62.354694}  
Iteration 3:{space 3}log pseudolikelihood = {res:-62.348999}  
Iteration 4:{space 3}log pseudolikelihood = {res:-62.348996}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:274}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:15.21}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0016}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-62.348996}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0968}

{txt}{ralign 86:(Std. err. adjusted for {res:21} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.4914074{col 34}{space 2} .4595187{col 45}{space 1}   -1.07{col 54}{space 3}0.285{col 62}{space 4}-1.392048{col 75}{space 3} .4092327
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}  .171314{col 34}{space 2} .1311172{col 45}{space 1}    1.31{col 54}{space 3}0.191{col 62}{space 4} -.085671{col 75}{space 3}  .428299
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -1.28216{col 34}{space 2} .3605393{col 45}{space 1}   -3.56{col 54}{space 3}0.000{col 62}{space 4}-1.988804{col 75}{space 3}-.5755163
{txt}{space 15}_cons {c |}{col 22}{res}{space 2} -1.70061{col 34}{space 2} .3237881{col 45}{space 1}   -5.25{col 54}{space 3}0.000{col 62}{space 4}-2.335223{col 75}{space 3}-1.065997
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1969, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     4,766
                                                {txt}F(3, 137)         =  {res}     2.66
                                                {txt}Prob > F          = {res}    0.0508
                                                {txt}R-squared         = {res}    0.0021
                                                {txt}Root MSE          =    {res} .09818

{txt}{ralign 86:(Std. err. adjusted for {res:138} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0038326{col 34}{space 2} .0032048{col 45}{space 1}    1.20{col 54}{space 3}0.234{col 62}{space 4}-.0025046{col 75}{space 3} .0101699
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0023607{col 34}{space 2} .0014988{col 45}{space 1}   -1.58{col 54}{space 3}0.118{col 62}{space 4}-.0053245{col 75}{space 3} .0006031
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0022847{col 34}{space 2} .0014774{col 45}{space 1}    1.55{col 54}{space 3}0.124{col 62}{space 4}-.0006367{col 75}{space 3} .0052062
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-7.52e-06{col 34}{space 2} .0035429{col 45}{space 1}   -0.00{col 54}{space 3}0.998{col 62}{space 4}-.0070133{col 75}{space 3} .0069983
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. *** APPENDIX TABLE 12***
.                 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>=1976, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-63.993221}  
Iteration 1:{space 3}log pseudolikelihood = {res:-63.204627}  
Iteration 2:{space 3}log pseudolikelihood = {res:-63.154653}  
Iteration 3:{space 3}log pseudolikelihood = {res:-63.154554}  
Iteration 4:{space 3}log pseudolikelihood = {res:-63.154554}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,059}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:1.86}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.6027}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-63.154554}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0131}

{txt}{ralign 86:(Std. err. adjusted for {res:135} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .4544728{col 34}{space 2} .4895864{col 45}{space 1}    0.93{col 54}{space 3}0.353{col 62}{space 4}-.5050989{col 75}{space 3} 1.414045
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.3437517{col 34}{space 2} .3478621{col 45}{space 1}   -0.99{col 54}{space 3}0.323{col 62}{space 4}-1.025549{col 75}{space 3} .3380455
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0197931{col 34}{space 2} .2771578{col 45}{space 1}    0.07{col 54}{space 3}0.943{col 62}{space 4}-.5234263{col 75}{space 3} .5630125
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.906189{col 34}{space 2} .6144824{col 45}{space 1}   -9.61{col 54}{space 3}0.000{col 62}{space 4}-7.110552{col 75}{space 3}-4.701826
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>=1976, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-51.658314}  
Iteration 1:{space 3}log pseudolikelihood = {res:-49.355544}  
Iteration 2:{space 3}log pseudolikelihood = {res:-48.570181}  
Iteration 3:{space 3}log pseudolikelihood = {res:-48.548615}  
Iteration 4:{space 3}log pseudolikelihood = {res:-48.548513}  
Iteration 5:{space 3}log pseudolikelihood = {res:-48.548513}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,132}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:14.81}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0020}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-48.548513}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0602}

{txt}{ralign 86:(Std. err. adjusted for {res:135} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .7031312{col 34}{space 2} .5502543{col 45}{space 1}    1.28{col 54}{space 3}0.201{col 62}{space 4}-.3753473{col 75}{space 3}  1.78161
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.077971{col 34}{space 2} .4105925{col 45}{space 1}   -2.63{col 54}{space 3}0.009{col 62}{space 4}-1.882718{col 75}{space 3}-.2732249
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0601004{col 34}{space 2} .4611335{col 45}{space 1}    0.13{col 54}{space 3}0.896{col 62}{space 4}-.8437047{col 75}{space 3} .9639055
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.779767{col 34}{space 2} .8755908{col 45}{space 1}   -6.60{col 54}{space 3}0.000{col 62}{space 4}-7.495893{col 75}{space 3} -4.06364
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>=1976, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res: -51.52998}  
Iteration 1:{space 3}log pseudolikelihood = {res:-50.357738}  
Iteration 2:{space 3}log pseudolikelihood = {res:-50.175374}  
Iteration 3:{space 3}log pseudolikelihood = {res:-50.175167}  
Iteration 4:{space 3}log pseudolikelihood = {res:-50.175167}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,057}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:4.48}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.2140}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-50.175167}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0263}

{txt}{ralign 86:(Std. err. adjusted for {res:134} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .9443894{col 34}{space 2} .5147377{col 45}{space 1}    1.83{col 54}{space 3}0.067{col 62}{space 4}-.0644781{col 75}{space 3} 1.953257
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.3490887{col 34}{space 2} .4734609{col 45}{space 1}   -0.74{col 54}{space 3}0.461{col 62}{space 4}-1.277055{col 75}{space 3} .5788775
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0654622{col 34}{space 2} .3612681{col 45}{space 1}    0.18{col 54}{space 3}0.856{col 62}{space 4}-.6426102{col 75}{space 3} .7735347
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.507685{col 34}{space 2} .8146238{col 45}{space 1}   -7.99{col 54}{space 3}0.000{col 62}{space 4}-8.104318{col 75}{space 3}-4.911052
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1976 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-93.600689}  
Iteration 1:{space 3}log pseudolikelihood = {res:-90.565138}  
Iteration 2:{space 3}log pseudolikelihood = {res:-90.000048}  
Iteration 3:{space 3}log pseudolikelihood = {res:-89.992729}  
Iteration 4:{space 3}log pseudolikelihood = {res: -89.99272}  
Iteration 5:{space 3}log pseudolikelihood = {res: -89.99272}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,132}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:14.78}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0020}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-89.99272}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.0385}

{txt}{ralign 86:(Std. err. adjusted for {res:135} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .7683737{col 34}{space 2}  .386627{col 45}{space 1}    1.99{col 54}{space 3}0.047{col 62}{space 4} .0105986{col 75}{space 3} 1.526149
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.6333937{col 34}{space 2} .3576019{col 45}{space 1}   -1.77{col 54}{space 3}0.077{col 62}{space 4} -1.33428{col 75}{space 3} .0674931
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0345642{col 34}{space 2} .2576935{col 45}{space 1}    0.13{col 54}{space 3}0.893{col 62}{space 4}-.4705057{col 75}{space 3} .5396342
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.411034{col 34}{space 2} .6250023{col 45}{space 1}   -8.66{col 54}{space 3}0.000{col 62}{space 4}-6.636016{col 75}{space 3}-4.186052
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>=1976, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-53.276893}  
Iteration 1:{space 3}log pseudolikelihood = {res:-47.677214}  
Iteration 2:{space 3}log pseudolikelihood = {res:-46.295241}  
Iteration 3:{space 3}log pseudolikelihood = {res:-46.267625}  
Iteration 4:{space 3}log pseudolikelihood = {res:  -46.2676}  
Iteration 5:{space 3}log pseudolikelihood = {res:  -46.2676}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:200}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:9.78}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0205}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 8:-46.2676}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1316}

{txt}{ralign 86:(Std. err. adjusted for {res:16} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.7265268{col 34}{space 2} .5107626{col 45}{space 1}   -1.42{col 54}{space 3}0.155{col 62}{space 4}-1.727603{col 75}{space 3} .2745495
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .2221245{col 34}{space 2} .1487708{col 45}{space 1}    1.49{col 54}{space 3}0.135{col 62}{space 4}-.0694609{col 75}{space 3} .5137099
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -1.55213{col 34}{space 2} .5177944{col 45}{space 1}   -3.00{col 54}{space 3}0.003{col 62}{space 4}-2.566989{col 75}{space 3}-.5372721
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-1.629547{col 34}{space 2} .3083764{col 45}{space 1}   -5.28{col 54}{space 3}0.000{col 62}{space 4}-2.233954{col 75}{space 3}-1.025141
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>=1976, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-53.276893}  
Iteration 1:{space 3}log pseudolikelihood = {res:-47.677214}  
Iteration 2:{space 3}log pseudolikelihood = {res:-46.295241}  
Iteration 3:{space 3}log pseudolikelihood = {res:-46.267625}  
Iteration 4:{space 3}log pseudolikelihood = {res:  -46.2676}  
Iteration 5:{space 3}log pseudolikelihood = {res:  -46.2676}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:200}
{txt}{col 57}{lalign 13:Wald chi2({res:3})}{col 70} = {res}{ralign 6:9.78}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0205}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 8:-46.2676}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1316}

{txt}{ralign 86:(Std. err. adjusted for {res:16} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.7265268{col 34}{space 2} .5107626{col 45}{space 1}   -1.42{col 54}{space 3}0.155{col 62}{space 4}-1.727603{col 75}{space 3} .2745495
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .2221245{col 34}{space 2} .1487708{col 45}{space 1}    1.49{col 54}{space 3}0.135{col 62}{space 4}-.0694609{col 75}{space 3} .5137099
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} -1.55213{col 34}{space 2} .5177944{col 45}{space 1}   -3.00{col 54}{space 3}0.003{col 62}{space 4}-2.566989{col 75}{space 3}-.5372721
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-1.629547{col 34}{space 2} .3083764{col 45}{space 1}   -5.28{col 54}{space 3}0.000{col 62}{space 4}-2.233954{col 75}{space 3}-1.025141
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  if aec2==1 & year_a>=1976, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     4,266
                                                {txt}F(3, 135)         =  {res}     2.91
                                                {txt}Prob > F          = {res}    0.0371
                                                {txt}R-squared         = {res}    0.0016
                                                {txt}Root MSE          =    {res} .08521

{txt}{ralign 86:(Std. err. adjusted for {res:136} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0005493{col 34}{space 2} .0025546{col 45}{space 1}    0.22{col 54}{space 3}0.830{col 62}{space 4}-.0045029{col 75}{space 3} .0056015
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0010465{col 34}{space 2} .0014291{col 45}{space 1}   -0.73{col 54}{space 3}0.465{col 62}{space 4}-.0038729{col 75}{space 3} .0017799
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0031217{col 34}{space 2} .0015149{col 45}{space 1}    2.06{col 54}{space 3}0.041{col 62}{space 4} .0001257{col 75}{space 3} .0061177
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-.0022124{col 34}{space 2} .0037439{col 45}{space 1}   -0.59{col 54}{space 3}0.556{col 62}{space 4}-.0096168{col 75}{space 3} .0051919
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
. 
. 
. ***APPENDIX TABLE 13*** 
. 
. *Program (Column 1)
. logit audience_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & (audience_on==0 | (audience_on==1 & audience_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-150.32008}  
Iteration 1:{space 3}log pseudolikelihood = {res:-137.80257}  
Iteration 2:{space 3}log pseudolikelihood = {res:-126.68669}  
Iteration 3:{space 3}log pseudolikelihood = {res:-125.71758}  
Iteration 4:{space 3}log pseudolikelihood = {res:-125.63817}  
Iteration 5:{space 3}log pseudolikelihood = {res:-125.63802}  
Iteration 6:{space 3}log pseudolikelihood = {res:-125.63802}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,647}
{txt}{col 57}{lalign 13:Wald chi2({res:6})}{col 70} = {res}{ralign 6:49.20}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-125.63802}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1642}

{txt}{ralign 86:(Std. err. adjusted for {res:138} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .8775451{col 34}{space 2} .3082468{col 45}{space 1}    2.85{col 54}{space 3}0.004{col 62}{space 4} .2733925{col 75}{space 3} 1.481698
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.056376{col 34}{space 2} .3410115{col 45}{space 1}   -3.10{col 54}{space 3}0.002{col 62}{space 4}-1.724746{col 75}{space 3}-.3880054
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0396853{col 34}{space 2} .2609969{col 45}{space 1}    0.15{col 54}{space 3}0.879{col 62}{space 4}-.4718593{col 75}{space 3} .5512299
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2} .1168331{col 34}{space 2} .0291847{col 45}{space 1}    4.00{col 54}{space 3}0.000{col 62}{space 4} .0596321{col 75}{space 3}  .174034
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} 1.486063{col 34}{space 2} .4568665{col 45}{space 1}    3.25{col 54}{space 3}0.001{col 62}{space 4} .5906213{col 75}{space 3} 2.381505
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2} 1.200342{col 34}{space 2}   1.1432{col 45}{space 1}    1.05{col 54}{space 3}0.294{col 62}{space 4}-1.040288{col 75}{space 3} 3.440973
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.134297{col 34}{space 2} .5851785{col 45}{space 1}  -10.48{col 54}{space 3}0.000{col 62}{space 4}-7.281226{col 75}{space 3}-4.987368
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Pursue (Column 2)
. logit pursue_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5   MID5 totrival cowmaj if aec2==1 & (audience_on==0 | audience_explore ==1 | (audience_on==1 & pursue_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-107.51186}  
Iteration 1:{space 3}log pseudolikelihood = {res:-95.745849}  
Iteration 2:{space 3}log pseudolikelihood = {res: -85.86492}  
Iteration 3:{space 3}log pseudolikelihood = {res:-83.127031}  
Iteration 4:{space 3}log pseudolikelihood = {res:-82.736461}  
Iteration 5:{space 3}log pseudolikelihood = {res:-82.728977}  
Iteration 6:{space 3}log pseudolikelihood = {res: -82.72896}  
Iteration 7:{space 3}log pseudolikelihood = {res: -82.72896}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,884}
{txt}{col 57}{lalign 13:Wald chi2({res:6})}{col 70} = {res}{ralign 6:77.63}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 9:-82.72896}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.2305}

{txt}{ralign 86:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        pursue_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .4984767{col 34}{space 2} .4304689{col 45}{space 1}    1.16{col 54}{space 3}0.247{col 62}{space 4}-.3452269{col 75}{space 3}  1.34218
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.439529{col 34}{space 2} .6976126{col 45}{space 1}   -2.06{col 54}{space 3}0.039{col 62}{space 4}-2.806824{col 75}{space 3}-.0722331
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .3655558{col 34}{space 2} .3511253{col 45}{space 1}    1.04{col 54}{space 3}0.298{col 62}{space 4}-.3226373{col 75}{space 3} 1.053749
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2} .0874758{col 34}{space 2} .0113701{col 45}{space 1}    7.69{col 54}{space 3}0.000{col 62}{space 4} .0651907{col 75}{space 3} .1097608
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} 1.939445{col 34}{space 2} .6006998{col 45}{space 1}    3.23{col 54}{space 3}0.001{col 62}{space 4} .7620948{col 75}{space 3} 3.116795
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2} 1.741336{col 34}{space 2} .7668475{col 45}{space 1}    2.27{col 54}{space 3}0.023{col 62}{space 4} .2383421{col 75}{space 3} 3.244329
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.634488{col 34}{space 2} .6343952{col 45}{space 1}  -10.46{col 54}{space 3}0.000{col 62}{space 4} -7.87788{col 75}{space 3}-5.391096
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Explore (Column 3)
. logit explore_start  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & (audience_on==0 | (audience_on==1 & explore_start==1)) & year_a>1945, vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-134.32743}  
Iteration 1:{space 3}log pseudolikelihood = {res:-120.31871}  
Iteration 2:{space 3}log pseudolikelihood = {res:-110.49559}  
Iteration 3:{space 3}log pseudolikelihood = {res:-109.45751}  
Iteration 4:{space 3}log pseudolikelihood = {res:-109.29161}  
Iteration 5:{space 3}log pseudolikelihood = {res:-109.29121}  
Iteration 6:{space 3}log pseudolikelihood = {res:-109.29121}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,644}
{txt}{col 57}{lalign 13:Wald chi2({res:6})}{col 70} = {res}{ralign 6:40.95}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-109.29121}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1864}

{txt}{ralign 86:(Std. err. adjusted for {res:139} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       explore_start{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} 1.145618{col 34}{space 2} .3386877{col 45}{space 1}    3.38{col 54}{space 3}0.001{col 62}{space 4} .4818022{col 75}{space 3} 1.809434
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.153935{col 34}{space 2} .3902478{col 45}{space 1}   -2.96{col 54}{space 3}0.003{col 62}{space 4}-1.918806{col 75}{space 3}-.3890632
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.0336962{col 34}{space 2} .3475327{col 45}{space 1}   -0.10{col 54}{space 3}0.923{col 62}{space 4}-.7148479{col 75}{space 3} .6474554
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2} .1274342{col 34}{space 2} .0282321{col 45}{space 1}    4.51{col 54}{space 3}0.000{col 62}{space 4} .0721003{col 75}{space 3}  .182768
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} 1.220425{col 34}{space 2}  .606723{col 45}{space 1}    2.01{col 54}{space 3}0.044{col 62}{space 4} .0312696{col 75}{space 3}  2.40958
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2} 1.279212{col 34}{space 2} 1.042594{col 45}{space 1}    1.23{col 54}{space 3}0.220{col 62}{space 4}-.7642351{col 75}{space 3} 3.322659
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-6.541316{col 34}{space 2} .6849811{col 45}{space 1}   -9.55{col 54}{space 3}0.000{col 62}{space 4}-7.883854{col 75}{space 3}-5.198778
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * Accelerate (Column 4)
. logit audience_accel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & year_a>1945 & (audience_pursue ==0 | pursue_start==1), vce(cluster audience)

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-212.63423}  
Iteration 1:{space 3}log pseudolikelihood = {res: -210.0059}  
Iteration 2:{space 3}log pseudolikelihood = {res:-176.87775}  
Iteration 3:{space 3}log pseudolikelihood = {res:-173.51853}  
Iteration 4:{space 3}log pseudolikelihood = {res:-173.32202}  
Iteration 5:{space 3}log pseudolikelihood = {res:-173.32096}  
Iteration 6:{space 3}log pseudolikelihood = {res:-173.32096}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:4,884}
{txt}{col 57}{lalign 13:Wald chi2({res:6})}{col 70} = {res}{ralign 6:68.96}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-173.32096}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1849}

{txt}{ralign 86:(Std. err. adjusted for {res:141} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_accel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .7319369{col 34}{space 2} .2881836{col 45}{space 1}    2.54{col 54}{space 3}0.011{col 62}{space 4} .1671075{col 75}{space 3} 1.296766
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-1.133268{col 34}{space 2} .3778874{col 45}{space 1}   -3.00{col 54}{space 3}0.003{col 62}{space 4}-1.873914{col 75}{space 3}-.3926227
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .1593735{col 34}{space 2} .2291065{col 45}{space 1}    0.70{col 54}{space 3}0.487{col 62}{space 4} -.289667{col 75}{space 3}  .608414
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2} .0768446{col 34}{space 2} .0109338{col 45}{space 1}    7.03{col 54}{space 3}0.000{col 62}{space 4} .0554148{col 75}{space 3} .0982744
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} 1.527547{col 34}{space 2} .4572092{col 45}{space 1}    3.34{col 54}{space 3}0.001{col 62}{space 4} .6314332{col 75}{space 3}  2.42366
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2} 1.476093{col 34}{space 2} .7383238{col 45}{space 1}    2.00{col 54}{space 3}0.046{col 62}{space 4} .0290047{col 75}{space 3} 2.923181
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-5.685787{col 34}{space 2}  .486451{col 45}{space 1}  -11.69{col 54}{space 3}0.000{col 62}{space 4}-6.639214{col 75}{space 3}-4.732361
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Decelerate (Column 5)
. logit audience_decel  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" &audience_on==1 & year_a>1945, vce(cluster audience)

{txt}note: {bf:cowmaj} != 0 predicts failure perfectly;
      {bf:cowmaj} omitted and 32 obs not used.

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-81.943692}  
Iteration 1:{space 3}log pseudolikelihood = {res:-76.909307}  
Iteration 2:{space 3}log pseudolikelihood = {res:-73.640222}  
Iteration 3:{space 3}log pseudolikelihood = {res:-73.612879}  
Iteration 4:{space 3}log pseudolikelihood = {res:-73.612797}  
Iteration 5:{space 3}log pseudolikelihood = {res:-73.612797}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:393}
{txt}{col 57}{lalign 13:Wald chi2({res:5})}{col 70} = {res}{ralign 6:32.63}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-73.612797}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1017}

{txt}{ralign 86:(Std. err. adjusted for {res:24} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}      audience_decel{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-.9733203{col 34}{space 2} .3180884{col 45}{space 1}   -3.06{col 54}{space 3}0.002{col 62}{space 4}-1.596762{col 75}{space 3}-.3498784
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3256331{col 34}{space 2} .1151426{col 45}{space 1}    2.83{col 54}{space 3}0.005{col 62}{space 4} .0999579{col 75}{space 3} .5513084
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.9401682{col 34}{space 2} .4332822{col 45}{space 1}   -2.17{col 54}{space 3}0.030{col 62}{space 4}-1.789386{col 75}{space 3}-.0909506
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2}-.0595142{col 34}{space 2} .0285177{col 45}{space 1}   -2.09{col 54}{space 3}0.037{col 62}{space 4}-.1154078{col 75}{space 3}-.0036206
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} .2521109{col 34}{space 2} .6467791{col 45}{space 1}    0.39{col 54}{space 3}0.697{col 62}{space 4}-1.015553{col 75}{space 3} 1.519775
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2}        0{col 34}{txt}  (omitted)
{space 15}_cons {c |}{col 22}{res}{space 2}-1.596598{col 34}{space 2} .4189191{col 45}{space 1}   -3.81{col 54}{space 3}0.000{col 62}{space 4}-2.417664{col 75}{space 3}-.7755319
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * End (Column 6)
. logit audience_end  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & resp_a!="attack" & resp_a!="tolerate1" & audience_on==1 & year_a>1945, vce(cluster audience)

{txt}note: {bf:cowmaj} != 0 predicts failure perfectly;
      {bf:cowmaj} omitted and 32 obs not used.

{res}{txt}Iteration 0:{space 3}log pseudolikelihood = {res:-79.043781}  
Iteration 1:{space 3}log pseudolikelihood = {res:-74.254042}  
Iteration 2:{space 3}log pseudolikelihood = {res: -70.05776}  
Iteration 3:{space 3}log pseudolikelihood = {res:-69.979977}  
Iteration 4:{space 3}log pseudolikelihood = {res:-69.979888}  
Iteration 5:{space 3}log pseudolikelihood = {res:-69.979888}  
{res}
{txt}{col 1}Logistic regression{col 57}{lalign 13:Number of obs}{col 70} = {res}{ralign 6:393}
{txt}{col 57}{lalign 13:Wald chi2({res:5})}{col 70} = {res}{ralign 6:32.41}
{txt}{col 57}{lalign 13:Prob > chi2}{col 70} = {res}{ralign 6:0.0000}
{txt}{col 1}{lalign 20:Log pseudolikelihood}{col 21} = {res}{ralign 10:-69.979888}{txt}{col 57}{lalign 13:Pseudo R2}{col 70} = {res}{ralign 6:0.1147}

{txt}{ralign 86:(Std. err. adjusted for {res:24} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}        audience_end{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      z{col 54}   P>|z|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2}-1.028329{col 34}{space 2} .3458957{col 45}{space 1}   -2.97{col 54}{space 3}0.003{col 62}{space 4}-1.706272{col 75}{space 3}-.3503862
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2} .3515039{col 34}{space 2} .1155141{col 45}{space 1}    3.04{col 54}{space 3}0.002{col 62}{space 4} .1251004{col 75}{space 3} .5779074
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2}-.8941406{col 34}{space 2} .4434281{col 45}{space 1}   -2.02{col 54}{space 3}0.044{col 62}{space 4}-1.763244{col 75}{space 3}-.0250376
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2}-.0810576{col 34}{space 2} .0364927{col 45}{space 1}   -2.22{col 54}{space 3}0.026{col 62}{space 4}-.1525819{col 75}{space 3}-.0095333
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} .4545314{col 34}{space 2} .5895251{col 45}{space 1}    0.77{col 54}{space 3}0.441{col 62}{space 4}-.7009165{col 75}{space 3} 1.609979
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2}        0{col 34}{txt}  (omitted)
{space 15}_cons {c |}{col 22}{res}{space 2}-1.579479{col 34}{space 2} .4342365{col 45}{space 1}   -3.64{col 54}{space 3}0.000{col 62}{space 4}-2.430567{col 75}{space 3}-.7283907
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. *Status Change (Column 7)
. reg status_change  tol_end_total_full_5 Nat_total_full_5 dll_total_full_5  MID5 totrival cowmaj if aec2==1 & year_a>1945, vce(cluster audience)

{txt}Linear regression                               Number of obs     = {res}     5,070
                                                {txt}F(6, 142)         =  {res}     4.32
                                                {txt}Prob > F          = {res}    0.0005
                                                {txt}R-squared         = {res}    0.0083
                                                {txt}Root MSE          =    {res} .10747

{txt}{ralign 86:(Std. err. adjusted for {res:143} clusters in {res:audience})}
{hline 21}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 22}{c |}{col 34}    Robust
{col 1}       status_change{col 22}{c |} Coefficient{col 34}  std. err.{col 46}      t{col 54}   P>|t|{col 62}     [95% con{col 75}f. interval]
{hline 21}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
tol_end_total_full_5 {c |}{col 22}{res}{space 2} .0055647{col 34}{space 2} .0030593{col 45}{space 1}    1.82{col 54}{space 3}0.071{col 62}{space 4}-.0004829{col 75}{space 3} .0116124
{txt}{space 4}Nat_total_full_5 {c |}{col 22}{res}{space 2}-.0030348{col 34}{space 2} .0013556{col 45}{space 1}   -2.24{col 54}{space 3}0.027{col 62}{space 4}-.0057146{col 75}{space 3} -.000355
{txt}{space 4}dll_total_full_5 {c |}{col 22}{res}{space 2} .0012905{col 34}{space 2} .0014197{col 45}{space 1}    0.91{col 54}{space 3}0.365{col 62}{space 4}-.0015159{col 75}{space 3} .0040969
{txt}{space 16}MID5 {c |}{col 22}{res}{space 2} .0012104{col 34}{space 2} .0005551{col 45}{space 1}    2.18{col 54}{space 3}0.031{col 62}{space 4}  .000113{col 75}{space 3} .0023077
{txt}{space 12}totrival {c |}{col 22}{res}{space 2} .0117017{col 34}{space 2} .0115724{col 45}{space 1}    1.01{col 54}{space 3}0.314{col 62}{space 4}-.0111747{col 75}{space 3} .0345781
{txt}{space 14}cowmaj {c |}{col 22}{res}{space 2} .0101714{col 34}{space 2} .0207086{col 45}{space 1}    0.49{col 54}{space 3}0.624{col 62}{space 4}-.0307656{col 75}{space 3} .0511084
{txt}{space 15}_cons {c |}{col 22}{res}{space 2}-.0032507{col 34}{space 2} .0045916{col 45}{space 1}   -0.71{col 54}{space 3}0.480{col 62}{space 4}-.0123274{col 75}{space 3}  .005826
{txt}{hline 21}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{res}{txt}
{com}. 
{txt}end of do-file

{com}. log close
      {txt}name:  {res}<unnamed>
       {txt}log:  {res}C:\Users\MB\Desktop\JOP_Replicate\ReplicationUpload\LogforRegressionModels.smcl
  {txt}log type:  {res}smcl
 {txt}closed on:  {res} 9 Feb 2023, 18:58:29
{txt}{.-}
{smcl}
{txt}{sf}{ul off}